SourceForge.net Logo





An fserve for XChat




2005-11-10 :: The End

Yeah, it's true... This is the end of TuxServe. Well, it actually ended a long time ago, but I never bothered updating the site. I will not be developing TuxServe any further. No more bug fixes and no more additions.

What do you use instead? Try Obsidian+ - it is an updated and modified version of the old Obsidian fserve. This fserve features most of TuxServe's features and, because it's written in perl, it should be much more portable. Having said that, I have not actually used Obsidian+ - so judge for yourself.

A few final, important notes about TuxServe: This site will remain, the project is still GPLed, it is full of bugs (which I will not be fixing), it does NOT work on OS X, I won't be accepting any further patches for it and I am no-longer providing support for it (since most of the problems relate to things I won't be fixing).

Thanks to everyone who supported the project; from the testers and bug reporters, to those who submitted patches. A list of thanks is available here. You guys and gals made the project a load of fun and I really appreciate all the help you gave me.

So long and thanks for all the fish.


Aypok...

2005-07-25 :: One Angry Dwarf And 200 Solemn Faces!

Greeting TuxServians, I come bearing more good news :)

First up is the porting of my Welcome Back plugin to Windows! Credit for this feat goes to Jory Stone - so thanks a lot dude :) It's nice to see people making use of the stuff I write, and even more cool to see the GPL in action (as said by Eno_). Now all the windows using X-Chat peoples can use the blinding WB plugin.

The Win32 pack can be downloaded from Jory's site. I should probably state that I haven't actually tested this out for myself (I have no Windows boxen) - let me know how it goes.

There will also be a new patch for the latest version of TuxServe coming out soon to fix a few of the bugs. I can't give a date, but it should be soon. Within a week or 2. Maybe... Check back next week to see if I've released it or not. If I haven't, feel free to come and bug me.

Enjoy.


Aypok...

2005-06-04 :: TuxServe 0.0.12.4alpha Released!

It's here! Finally, a new release! Yay. Party time.

Hands up: how many of you thought I wouldn't actually get around to releasing this? All of you? Have faith! _@_" <-- that's a lowly snail, yet it has faith! Be like the snail. Try not to leave that slimy stuff everywhere, though - that's annoying to clean up...

Anyhoo, back to the release. This release contains a great many bug fixes over previous versions, so it should be even more stable than before. There aren't many new features in this new release, but there may be one or two. This version *should* also support 64bit x86 machines properly - but that is currently still "theory", as it has not been tested. Let me know how it goes. Let me know of any bugs/problems, too - the usual place.

Get it here.

Enjoy.


Aypok...

2005-06-02 :: Wallace And The Anti-Wallace.

Minor updates to the code have been made and it is hopefully now fully compatible with x86 64bit machines. Thanks to 'Lord Drantin' for pointing out the problems with TuxServe on 64bit machines. This new stuff is on the CVS repository. Please let me know if it doesn't work or if you find any bugs.

There will hopefully be a new release out within a day or 2. It's been way too long since the last one. Check back tomorrow or the day after to see if it's out yet :)

Remember that XMMS plugin for X-Chat I mentioned I'd make available at xmas? Yeah, well I finally released it. There is no nice page for it yet, but the 'README' file should explain everything you need to know. If not, you know where to find me. It can be obtained from HERE.

While you're at it, check out my new site: www.aypok.co.uk :)

So, until the new release: enjoy.


Aypok...

2005-05-24 :: Octothorp Vs. Walls Of Ham.

Finally! An update. It's been a long time coming, but it's here. Although it's not an actual release - just some updated source. This source was supposed to uploaded over a week ago, but as I was working on the code, my phone line got killed by some strong winds and I was offline for a while.

But I'm back, and so is the new code. Pretty much ALL of the new code is by "pabst", so I must give him a great deal of thanks, once again, for his efforts. Thank you pabst - you really are a great help to this project.

This code may have a few "odd" bugs in it, however (which are my fault). When appling pabst's code to my own (which differed slightly from the CVS code) there were some parts that had to be patched manually. I found and fixed (with the help of pabst) 2 of these bugs, but there may be more out there. Please let me know if you come across any.

You know what this project needs? A new release. It really does. I'd love to make that happen sometime this week, but I'm not sure if that's possible. The main reason being that my testing platform (the box I used to test TuxServe) is gone - without it, testing TuxServe a very slow and annoying process.

I might try adding in some of those long-promised features or back-port those bug-fixes from 0.0.13, but without the ability to test the new code it may do more harm than good. We'll see what happens :)

For now; enjoy.


Aypok...

2005-04-02 :: Spectacular Paging Trials.

Just a quick update to let you all know that the CVS code has been updated. It was updated almost a week ago, but it was updated again today, so I thought I'd let you all know so you benefit from it :)

All of the new code is a collection of several patches written by the brilliant "pabst" - none of my code yet (I've still got many other things eating away at my time...). I've been playing around with the new code for a while and it all seems to work great (although I've not done any full tests as of yet). Thanks again to "pabst"!

Although this isn't a release, the code seems stable. If you find any bugs or experience any crashes, please let me know. This new code should also fix problems with opening and closing files, so I advise those of you whom had such problems to try the new code.

Update:: Okay, I should clarify something about the code, since the versioning and patching thing has gotten very confusing. There are 2 main code trees at the moment: the 0.0.12 tree and the 0.0.13 tree. Unless you checked out the CVS code after 0.0.12 was released, you won't have seen 0.0.13. I started work on it just after 0.0.12 was released and fixed many bugs and added some cool stuff in. But all subsequent patches (the ones I wrote and the ones i received) were against 0.0.12 - which is what I wanted, since 0.0.13 was slightly more unstable at the time.

Looking back on it, 0.0.13 has some stuff I want to bring back to the current TuxServe code tree, but doing so is a LOT of work. So, over the next few days I'm going to hopefully be applying the patches (by hand) to the 0.0.13 tree, to create the new code. If I manage to pull this off without screwing up anything, it'll be well worth it :)

The original plan was to scrap 0.0.13 and continue the 0.0.12 tree as version 0.0.14, but there is too much stuff in 0.0.13 to just throw it away. The CVS is currently version 0.0.14, which is 0.0.12 with some patches. This means that some of the cool stuff (excluding patches) mentioned here won't be available in the CVS code (eg, "auto_list").

Most of you wouldn't have known that, and most of you won't care. And I guess there was no point in me explaining it all :) But I thought I should clarify it incase anyone had been confused. If you weren't, you probably are now. My versioning schemes are great, aren't they? :D

For info on how to get the CVS code, click HERE. Enjoy.


Aypok...


2005-02-14 :: Cringing Nurglers Of Spon.

Hey hey, welcome to the first news in almost a month. Some important (meh) news to mention today, so on with the show:

I hate to do this, but... Mac OS X is no longer an "officially" supported platform in the TuxServe world. I know there is a huge number (comparatively) of OS X users who use TuxServe, but I'm still getting bug reports from OS X users. The bugs aren't simple things to fix, they're downright confusing, as they give no problem on other platforms (for instance; TuxServe will just fail to reopen its files). This, however, is only temporary! Full OS X support will be back in the future (see below for more info). I *will* still try and help you with running TuxServe on OS X, but I can't promise it'll work.

As you'll have noticed, as well as a lack of news updates, the code has been left alone a bit. I've not abandoned this project, no no! Far from it. But I've decided to take a small break from coding as a whole for a short while, to allow me to concentrate on some other things.

Anyhoo, now for the good news:

I have a very large (and quite radical) plan for TuxServe. It will take a while to accomplish, but it will be worth it for everyone. It will also make it far more portable (in theory) and have much better support for OS X. I'm not going to release any details of it yet - I want to get it at least part of the way off the ground before letting it out of the bag, as the saying goes :)

Until that's done, there will still be normal updates and patches. When I get back to coding. Enjoy.


Aypok...


2005-01-20 :: TuxServe 0.0.12.4 Patch.

As promised, here's the new patch for TuxServe. I won't go into too much detail here, as that's all explained in the included readme file. You can download it HERE. The main reason for this patch is to fix the crashing when a user leaves the fserve - but there are other changes, too.

Unlike previous patches, this is just a diff file. Well... 2 diff files... /me explains:

There are 2 patches included in the tar - you only need apply one of them, however. One is a patch from the basic TuxServe 0.0.12.2 pack, and one is from the patched version of that. There are more details in the readme. At the time of writing, this patch has not been ported forward to the CVS code. That should be done by tomorrow - maybe tonight.

Incase you're wondering why it skipped from 0.0.12.2 to 0.0.12.4: it didn't :) There is a 0.0.12.3, which is 0.0.12.2 with the first patch... Yeah, needlessly complicated. But it's not important.

UPDATE :: If you downloaded the patch before 19:00 GMT today, please redownload it. The original patch had a problem which has been corrected in the current version. Sorry for the inconvenience. Thanks to "fuzzybear" for spotting it!

Until the next update; enjoy.


Aypok...


2005-01-19 :: Dragon Typed Kettle Instructions.

Good news! (again!) There's a new patch for v0.0.12.2 scheduled to be released in the next day or 2. This fixes several things, but most importantly - the "crashing when user leaves the fserve" bug, which was introduced in the last patch. A full list of changes and improvements will be included when the patch is released. The new patch is based on code submitted by "pabst" - so a great deal of thanks is owed to him for his great work.

The new code isn't yet on the CVS (0.0.13), but I'll port the modifications forward after the patch is released. It seems like the 0.0.12.2 version is the "stable" branch at the moment, and the CVS version (0.0.13) is "unstable". This works out nicely, since 0.0.13 won't be usuable for a while (due to the large changes in the queues/sends system), while 0.0.12.2 is relatively stable (especially with this latest patch). So, if you're planning to submit any patches or new code, please do so against the latest patched release, rather than the CVS (Yes... I know that's a bit backwards from what usually happens).

I don't actually have a set date for the release of the patch - but I'm guessing it'll be done by tomorrow (2005-01-20), or the day after that at the latest.

Please keep the submissions coming in, and a huge amount of thanks to everyone who's helped out with this project :) Enjoy.


Aypok...


2005-01-08 :: Spleens Of Desire.

I've finally confirmed a few bugs that I asked about in the last news post. It seems that the "crashing when someone exits the fserve" bug in the patched version of TuxServe 0.0.12.2 has been experienced by more than one person. That's now an "officially known bug", and will hopefully be an "officially fixed bug" soon. It's still an unknown on the CVS version (but I'm guessing that it's broken too).

As for XChat Aqua (Mac OS X) crashing when you load TuxServe, there appears to be a simple and effective fix for this: After installing TuxServe, go to the dir you installed it in ("/Applications/x-chat_aqua/plugins/tuxserve/" by default) and delete the file "config.cfg". When you try reloading TuxServe, it shouldn't crash - although it will give an error saying that it cannot find the config file (duh). Just change the config settings to what you want, then save them out (with the command "/ts config save") - this will create a new config file, and should work fine when you reload TuxServe.

I don't know the exact reason that XChat Aqua crashes with TuxServe's default config - I assume it doesn't like something about the default config. I'll try and do a lot more testing of TuxServe on Mac OS X before the next release, to hopefully increase compatability and remove bugs like this.

In other news: The FAQ has also been updated a bit, mainly with the info above - nothing major. TuxServe 0.0.12.2 has had 250 downloads between 2004-12-08 and 2005-01-06 - it seems to be getting more popular all the time :)

Over the next week, TuxServe will undergo a rather large restructuring (code wise), as I'll be rewriting the queues and sends systems to work in a different way. While this will mean that the CVS will be largely unusable in that time, and the bug fixes will be few and far between (until it's done) - it will mean support for multiple IRC networks, priority queueing, resend on fail, etc. So it should be worth it :)

Update: It seems that someone has gone through a lot of effort and written a guide on how to install and configure TuxServe in Portuguese (I *think* that's the language - please excuse my ignorance). It can be found HERE - I'd like the thank the author for taking the time to write it :) I've added a link on the help page, too.

That's about all for today. Enjoy.


Aypok...


2005-01-05 :: Merry Everything.

Hey hey. First off, I'd like to apologise for the lack of updates (both to the site and the code) as of late - it's xmas, what can I say? :)

And, yes, I'm fully aware that it was the first Monday of the month 2 days ago - which is usually when I release a new version of TuxServe. The next version (0.0.13alpha) will be released on the first Monday of February. Which is a long way away, I know...

I didn't "forget" to release TuxServe 2 days back, I just decided that, due to the lack of improvements recently, there wasn't really a need to release a new version. I'm also going to go out on a limb and *promise* - yes, promise - that v0.0.13 alpha will have priority queueing ability. Yeah, yeah - it's not /that/ great... But I started work on it ages ago, so it's time that I finished it. I *should* (in theory) have enough time to add in "resend on fail" functionality - but I'm not going to promise that just yet, since there needs to be a largish rewrite of the queueing and sending functions for that to happen. This promise doesn't count if I can't get it working... I'll try my best :)

Oh yeah - as of 2005-01-01, the latest version of TuxServe has been downloaded 223 times! I'm actually surprised - and pleased, you understand :)

Reports of crashing: I've had 2 reports over the past few days that TuxServe crashes XChat Aqua (on Mac OS X) on loading it. I'm not too sure why. In both cases, however, they mentioned that they were loading it through a menu. If you have/had this problem, try loading TuxServe with the "/load" command, rather than through the menu. I'd appreciate it if you could let me know as to whether this fixes the problem or not.

I've had a report that the patched version of TuxServe 0.0.12.2 crashes when a user exits the fserve, or they time out (when "auto_close" is on). I've not been able to replicate this problem on my home systems, so I'd be grateful if anyone else who's had this problem could contact me (if anyone else has had this problem). My only advice is to either try the CVS code - but that may still be buggy, and uses the same code to exit a user from the fserve (the CVS code was backported for the bugfix). Or revert to the standard 0.0.12.2 release. I'll hopefully have fixed this soon.

Back to Mac OS X for a second - there is a known (?) problem where TuxServe will not remove any queues from the queues list, thusly resending the same queue over and over. Again, I've only had 1 report of this, so any input would be great. Apparently the text, when entered into TuxServe via its config commands, comes out garbled and random on Mac OS X machines. Not sure why... I've only had 1 report of it, though. It could just be a buffer overflow - if so, it'll be fixed on the CVS.

There will be no code updates for several days, since I've still got some other stuff to finish up first - but the code will floweth again by the start of next week. Which will hopefully bring many new features and bug fixes :)

As always; feel free to drop by on IRC and poke me, suggest features/ideas, report bugs, etc :) Enjoy.


Aypok...


2004-12-22 :: Salads Of The Cataclysm.

It's been a few days since the last update (not sure why) - but there is some news to report today. The main news is the release of a new bugfix pack.

The code on the CVS is mainly code cleaning and improvements of already implimented stuff. Like the new function that handles logging users into the fserve (same as before, but now it's been changed into 1 functions, instead of 2). Better support for Mac OS X (it will look for a config file in "/Applications/x-chat_aqua/plugins/tuxserve/" if it can't find the normal one). The displaying of the percentage of the file sent in the sends list should now be fixed. Pausing the fserve should no longer send the remaining queues (there was a bug where, if you paused TuxServe, it would send the queues). Better error handling on setting the context (not finished yet). The command "cd /" will now take the user to the root dir of the trigger, instead of refreshing the current dir.

The bugfix pack: This only fixes 2 minor bugs/problems with the 0.0.12.2 release. The displaying of the text "too many sends: x" has been fixed, and DCC chat windows will now autoclose (if it's enabled) when a user times out as well as when they quit properly. Go HERE to get it and for instructions on how to apply it.

I've also decided that today is a good day to release the WB (Welcome Back) plugin that I've been working on for the last few days. It just welcomes people back into the channel - nothing special. Click HERE to take a look. This is 1 of the xmas surprises that I mentioned a while back :)

Update: I've just fixed the "autolist" feature (CVS only) as requested by IceGSM :)

Enjoy.


Aypok...


2004-12-16 :: Across From Spell Check Castle.

Not much new news today, either. But there's some - as always :)

New on the CVS today are some bug fixes (for the previous CVS code - not for the release), a new help menu for the users browsing your fserve ("help commands"), and a new user-side command ("autolist on/off").

The CVS from a few days back had a rather large bug: if anyone used the tags to change dir, they would be transported to the root of the file system (/). This has hopefully been fixed - but I still advise that you don't use the latest CVS code for serving (I'm still using the latest stable release to serve - v0.0.12.2).

Someone pointed out that the automagical listing of directories could get annoying on very large fserves - and they're right. Due to this, I've added a new user-side command called "autolist", which accepts either "on" (which is default) or "off". It's pretty obvious what that does :) I've also added a new help command for users: "help commands" - this will give the users a list of all commands available to them, along with a brief description of each and an example of its usage.

That's all that's been added over the last few days. I was planning to do a lot more work on it - but I've been a bit low on sleep as of late, which impairs one's ability to code and think :) I'm still planning to have priority queueing and requeue-on-fail included and stable by the next release.

Until then; enjoy.


Aypok...


2004-12-11 :: Bridged From Footsteps.

Not much news today, but I feel like posting :)

Thanks to "gojyo`", a large bug has been fixed! It's the one where users get added to the list several times - if you have told TuxServe to serve on several channel, but you're not in them all, or if the same person is in more than 1 of your serving channels. That's been fixed and I'll make a bugfix pack if anyone requests it.

I've just finished going over the code and making it much safer - removing most possibilities of buffer overflows. This should mean; less bugs, less instability, and less security risks. I'm sure there are a few I've missed, but I'll fix them as I find them. It did break a lot of things when I did this, but I seem to have fixed them all - well, all the ones I've found.

All of these improvements are on the CVS, but I'd advise against using the CVS code for the next few days. I'm hoping to do some more large-ish changes, which could break large parts of TuxServe. If you want to try it anyway and see how it's going or to spot bugs, or if you just want to play with the code - please do. Just know that it is likely to contain some bugs.

That's all for today. More news soon. Enjoy.


Aypok...


2004-12-10 :: Splinters Of Toothpaste.

Hey all, I'm just here to ask for some help with tracking some possible bugs. If any of you can help me out with these, I'd really appreciate it.

I've had reports of TuxServe doing some strange things on Mac OS X - but only from a single person. It's not that I don't believe this person, it's just that I can't tell if it's a problem with their system (eg, that only affects them) or a bug that affects several/all other Mac OS X users. If you've used TuxServe on a Mac OS X box and have had; a) Trouble with setting the channels info - eg, the text you enter comes out randomised/garbled; b) It repeatedly sending the second queue over and over. The scrambled-data bug seems to go back several versions, but the second problem seems to only affect the latest version (maybe).

This next one affects Linux users. The problem is that TuxServe doesn't seem to log out users correctly, saying that they're already connected when they try to reconnect. I tried the server of the person reporting this problem, and it does exist on his system, but no one else has ever reported this to me.

So, if you've ever encountered either of these problems, please let me know. If you fixed them, even better - let me know how you did it :) Both of these problems won't reproduce themselves on my system. That doesn't mean they don't exist - I've had bugs reported to me that I couldn't reproduce, yet they were very real. Also, If you report a new bug to me, please note that it may take some time for me to find a fix. And don't be surprised if I've never before seen the bug you're reporting - if I had seen it, I'd have probably fixed it :)

A special note to "gojyo`": On re-reading what you wrote, it struck me that you might be using the option "auto_close". Although I seen no way that this could cause a problem such as the one you were having, try disabling it with "/ts config set auto_close 0". If that helps, please let me know and I'll try to figure out why it happens. If not... I guess I'll have to keep looking. Thanks for reporting the bug.

The 2 above bugs are the only 2 so far reported with the latest release - although it's far too early to say that it's bug free. Please keep the bug reports coming in - each bug report usually means another bug fixed, which in turn means it's better for you people :) Thanks again to all who've helped. Enjoy.


Aypok...


2004-12-08 :: TuxServe-0.0.12.2alpha Released!

The bugs (the major ones) that were in the last release, released just 2 days back, have now been fixed (/me does a little dance). This is a whole new release, as almost all files have been changed - so a bug fix pack would be pretty much the same as whole new release.

"What's with the version number?" Well, it's basically v0.0.12 - so the first version was called v0.0.12.1. But after I sent that out for testing, I added another feature (see below), so I incremented the version number. v0.0.12.1 does exist, yes - but only 1 person on the planet has a copy - I don't have a copy any more :)

"What's been fixed?" It will no longer say you have 1 queue whenever you load TuxServe, the problem where "sendnow"s caused the sends system to act weird, and the largest bug found so far; the sends/queues problem where some queues wouldn't get sent.

There have been 2 new features added, too: temporary file handling has been improved and there is now a new option called "temp_type" (which defaults to "1"). In previous versions, temporary files were created with the call "tmpfile()", but this seemed to cause a problem on at least 1 Mac OS X system. If "temp_type" is set to "1" it will continue to use this method - which is the recommended way. But, if TuxServe gives you errors with the temp file (for example: "ERROR: Cannot open temp file."), then set "temp_type" to "0" - this will make TuxServe use a normal text file for it's temporary needs. This temp file will be called "temp" and live in the TuxServe directory along with the config/sends/etc files.

The second new feature is more of an addition to a previous feature. When users use the "/ctcp your_nick my_queues" command, it will now tell them how many free queue slots they have, and how many free slots there are on the server in total. I just thought it'd be a cool addition :)

At the time of this release, I only know of 1 bug: the percentage reported by TuxServe when it lists the current sends (when a user browsing the fserve types "sends", or when you use the command "/ts csaq show_sends") will always be "0". This will be fixed at a later date.

If you find any more bugs, please let me know. Enjoy the new release.


Aypok...


2004-12-08 :: Splatter Noise.

Not long after I released the latest version (several hours, infact) I'd discovered 2 new bugs - 1 of which is quite annoying. I was going to post about them, but decided to try and fix them first. That failed.

It's now about 24 hours later, and I've spent another several hours trying to fix them, but to no avail. The 2 bugs I first mentioned actually seem to just be 1 bug, but causes 2 different problems. It sounds like a simple thing to fix - but it's proving difficult to track down.

The bug I'm talking about is that it will not send some queues - usually the last queue. The second bug is that the number of sends is reported wrong - it keeps getting higher and higher. It now seems that this could be the cause of the queues problem (as it thinks there are too many sends) - but I'm still not certain.

I actually noticed and fixed 2 other minor bugs earlier. The "send now" feature checked if they had free queue slots before sending a file - so, even if the file was eligible to be sent now, it would say "too many queues" if they had filled their queue slots. This has been fixed. Also, when starting TuxServe, it always said there was 1 queue - even if you'd set it to not resume files and/or the queues file was empty - that's been fixed too. There's been a new setting added to deal with how you want temp files to be handled (mainly for Mac OS X users), but I'll explain that another time.

Once again, I'm sorry that this release has these bugs - and I'm doubly sorry that I still haven't fixed them nearly 2 days later... I really hope to have them fixed sometime later today. I'll then release a bugfix pack, or a whole new release - or something. I'll work it out later :)

Enjoy.


Aypok...


2004-12-06 :: TuxServe-0.0.12alpha Released!

Here it is; the xmas version, I guess :) There are no xmassy treats inside - mainly because I only just thought of the idea. Even if I did, I'd probably create some crippling new bug - so it's best as it is :)

Anyhoo; back to business. This version, 0.0.12alpha, is mainly a bug fix release over 0.0.11alpha - but it does fix several important bugs. Some of the main points: Users can no-longer exceed the "user_sends" limit, fixed the bug where TuxServe ignored CTCP triggers if only set to use 1 channel, added the command "/ts csaq show_users" to list the currently connected users, if a queue can't be sent because the user hasn't got any free send slots, it'll try the next queue (until the end of the queues file), and last - but not least - vastly superior handling of users. In previous versions, if more than 1 user was browsing the fserve at a time, although it worked to an extent, TuxServe would screw up when one of them left - leaving one in a "zombie" fserve session (it's alive - but doesn't do anything).

All in all, it's definatly worth the upgrade. Remember: if you are upgrading, don't install it with the command "make install" - use "make upgrade". That will replace the binary, but keep your old settings and queues/sends/bans files.

As usual - there are no obvious bugs in this release. But remember that I say that almost every time :) This time, however, I mean it. It's been over a month since v0.0.11 was released - yet I only got 1 minor bug report (which has been fixed). This is a good sign, and it may mean that TuxServe is ready to go beta. We'll have to see about that, though.

So, go and download it. Please report all bugs and/or problems to me - even small ones - feel free to drop by on IRC and have a chat :) You may even get a copy of the soon-to-be-released XMMS plugin that I'm working on. Once again, thanks to everyone who's reported bugs and or supported me and TuxServe - you've been a great help. Enjoy the new release.


Aypok...


2004-12-05 :: Bottled Fan Blades.

Hi ho, merry peoples. It appears that, with help from "doublej", I've fixed the problem that some of you were having - TuxServe ignoring CTCP triggers. There are 2 ways to fix it: 1) as it affects you if you have only set 1 channel, set a second channel, or 2) download the CVS code. Or 3) wait until a whole new release is, uh, released... Tomorrow.

Tomorrow's (Monday, 2004-12-06) release is mainly a bug fix release over v0.0.11. It's just what's been living in the CVS for the last month - but I strongly recommend that you upgrade to the latest version if you're still on v0.0.11. If you're using the CVS code from as long ago as a month, and you use more than 1 channel, it's safe to stick with that (it's basically the same as the release) - but upgrading won't hurt :)

The new features for TuxServe will probably arrive sometime next year (eg, January), which is later than I'd hoped - but I've been side tracked with other things recently. The new features to which I'm refering are things like priority queuing (which, as I said before, is nearly done), re-send/queue on failed sends, and of course - a GTK2 config tool. The config tool won't be here by January, but the rest might :)

More news of the release, and the actual release pack it self, will be available tomorrow morning/noon (GMT) - so be sure to check back then. Enjoy.


Aypok...


2004-12-04 :: Trouble With A Capital Moo.

It turns out that the latest version of TuxServe isn't as bug-free as I'd thought. The problem is that some people (note "some") can't get CTCP triggers to work at all, even though public triggers work fine. The most peculiar thing about it is the timing - there have been 3 reports of it in the last 4 days - even though the code hasn't been modified for weeks. Also, there's a new version release planned for the coming Monday - which may be postponed until a fix is sorted. Again.

When this was first reported to me, I diagnosed it as a simple mis-configuration, but as I was helping him fix it, I realised that this was not the case. After ruling out certain other things (XChat itself was fine, as public triggers worked fine, and he could see the CTCP requests sent to him - but TuxServe just ignored them), I thought it might be a system specific problem (as it is still a problem I cannot reproduce on any of my systems - no matter what I try) and sent him a special "debugging" version that would have told me where the problem was. Unfortunatly, I never saw him again :(

The next day, someone else comes in with the same problem, and again yesterday. Alarm bells started to ring in my head and I decided that I should let you all know of a possible problem with TuxServe. But, disappointingly, I cannot see any problems in the code, and it seems like almost everyone has no such problem with TuxServe. Which, along with the timing, makes it even more curious.

I am currently working on a fix for this - or rather, a way to find where the actual problem lies. So stay tuned to this site for the next few days - or drop by on IRC - for news of a fix/new release. The new release is now planned for whenever this is fixed.

Once again, thanks to everyone who reported this to me :) Enjoy.


Aypok...


2004-11-23 :: A Barrel Of Fish Discs.

It's been about 2 and a half weeks since the last update (of the site and the code), so I though I'd post a little message to let you know everything's fine. Everything's fine.

"Are you crazy?!" Quite probably, but I've been distracted by other things recently. I have a new project in the works (more info when it's done) and I've been playing TFC a lot. Again :) Plus, I've not had a single bug report on the latest version. Yup - none at all! Seriously. That means that this version doesn't need too much fixing at the moment, and that I can take a short break. Don't worry; I'll be back to coding this in a week or 3.

Also, I've finally gotten around to linking to the "No Software Patents" site. I'm in the EU and feel very strongly about this matter - I've already contacted my MP and government officials about it. But, that may not be enough :( If you're in the EU, you'll know what I'm on about - and probably feel the same way as I, and most other Europeans, do. Please check out the links and help us out.

Enjoy.


Aypok...


2004-11-05 :: The Fuselage Message.

Happy bonfire/Guy Fawlks night! If you're celebrating it, that is. If you're staying in, like me, to waste hours on IRC and ignore the outside world - good luck, too.

Back to the topic at hand. The problems with the users system that I mentioned yesterday has now been fixed, it seems. I've tested it and had several users connected at once. They could all do what they wanted (within the limits you set) - even after others logged out and back in again. This fix is on the CVS.

And, uh... Yeah. That's all for today. Don't go blowing yourselves up tonight. Enjoy.


Aypok...


2004-11-04 :: Chubba Chubba Splin.

/me starts off today with a warning - make that 2 warnings.

"send now" is broken. In a rather stupid move, the latest release has this enabled by default. Guess I should have checked it out... Anyhoo, the fix is on the CVS - but a better idea would be to just disable send_now with: "/ts config set send_now 0". This bug will send EVERY file bigger than "send_now" to them instantly, ignoring the queues and the sends limit (a different bug to the known max_sends bug - which was fixed yesterday, and is also on the CVS repository).

The second warning is about multiple users on your fserve at once. In short; it doesn't work properly. "What do you mean? I've had more than 1 user browse my fserve at a time!" Yes, you can have many users logged in and using your fserve at once - and they can queue files and browse fine, etc. The problem occurs when a user logs out - it seems to screw up some stuff, which means the other users are left in a normal DCC chat session. They need to reconnect to continue. I'm working on a fix, but the earliest it'll be done is tomorrow. A "fix" for this is to limit "max_users" to 1. I'm sure most of you can live with letting just 1 user on at a time :)

Good news time: It seems that when I fixed the problem with logging into the fserve when TuxServe had an open dialog with them, I also sort of fixed the multiple server thing. You read that right (maybe - I don't know how you read it, so I'm guessing): TuxServe will respond to "!list" and trigger requests (CTCP and public) from any IRC server you're connected too - no matter which is currently focused. BUT, sending the files still seems to be a problem: if the person you're sending to is not on the currently focused server, it won't find them. This should be fixed within a day or 2 :)

The rest of the coding that went on today was just sorting out priority queueing (almost ready) and general fixes/updates. It's all on the CVS server, if you want to test it out. Enjoy.


Aypok...


2004-11-03 :: ...Then *HE* said, "Cheese?".

Howdy doody, people-bots. This morning turned out to be rather productive. I say "this morning", because my brain gave out pretty quickly due to a lack of sleep. I spent the rest of the day play chess on FICS, and OTTD.

Now, for a brief moment, I shall take us all on a journey to somewhere I like to call "reality". This is where most of the code changes took place (some code changes were made outside or reality, but they're not of much use to you). The changes have been lined up and identified as: Users can no-longer exceed the "user_sends" limit (eg, have more sends at once than they should do); If a queue can't be sent because the user hasn't got any free send slots, it'll try the next queue; It now checks to see if a completed/failed DCC send was from the fserve or not (if it wasn't, it won't send the next queue).

All of it is in the sends department, as you probably noticed. I also started work on implimentation of priority queuing for voiced/oped users. It's partly implimented, but it ignores them for now (other bits need to be added before it'll work). I didn't get a chance to test these out, but in thoery they should work fine (where've we heard that before?). I'll be testing them out soon enough, and fixing any problems found.

That's it for today - more to come tomorrow: hopefully some more work on priority queuing and maybe even an "Average Queue Time" thing, so users know how long they have to wait :) Until then: enjoy.


Aypok...


2004-11-01 :: TuxServe-0.0.11alpha Released!

Just as I promised :) This version has some cool "last minute" fixes that I did yesterday (thanks to Eno_ for spotting them!). So, if you want to skip all this crap, just go to the downloads page and grab yourself a copy. You can stay here and hear me ramble on about what's new in this version, if you wish. You do? Goody.

If you've been following the site and these wonderific news posts of mine, then you'll probably be familiar with most of the new things since the last version. The main points will now be recapped: "!voiceme" has been removed, compiles under GCC 2, public triggers work as well as CTCP ones, ban by host, better handling of DCC fails and timeouts, and loads of bug fixes (I mean LOADS). That's not all - I've saved the best for last :)

These so called "last minute" fixes that I mentioned are pretty important: "!list" will now work on ALL channels set to, not just the first - and (finally!), the bug where users can't re-login if you have a dialog window open with them has been fixed! Woo. I know that annoyed a lot of people - as well as me - so you'll be pleased with that.

I think I may be on the verge of fixing the bug where it will only respond to requests from channels/users on the currently focussed server - but that will have to wait until the next release - sorry. That's still a "known bug".

Talking of known bugs, there is some good news and bad news: The problem with forgetting to put a "/" on the end of the trigger_dir path is no more - it will automagically add one for you - and it works this time. But, the "user_sends" limit is not enforced in this version. The chances of this being a problem are slight, but you should be aware of it. To fix this would mean messing with the queues/sends system - but I didn't want to screw with them this close to a release and risk delaying it (not after last time). This'll hopefully be fixed very soon, and may be released as a bug-fix pack. It'll be on the CVS for sure. Keep checking the site for updates.

It also seems that this release is slightly smaller than the last one :) That's because all of the old (and unused) queues functions have been removed - as well as all the "!voiceme" related code.

I've just finished updating some of the help pages. The "Usage" section now contains all the commands available, and "Configuring" has been updated to include (and explain) the complete set of options. If I've forgotten anything (I usually do), let me know.

Anyhoo, installing and configuring TuxServe: if you've done this before, you'll probably be somewhat pleased to learn that it still works the same way. If you have a previous version of TuxServe installed and wish to use its settings, use the "make upgrade" option instead of "make install" (as always). If you're new to TuxServe, or want a refresher, check out the included "readme" file - or even better; check the help section. If you still have problems and/or questions, drop by on IRC (see the contact page) and let me know. Don't forget bugs! If you find bugs, please let me know :) I'll do my best to fix them as soon as possible.

Anyhoo - enjoy the release.


Aypok...


2004-10-29 :: Nothing Cleans Like THE SPON.

Welcome to another edition of "I was too busy doing other stuff to do much work on TuxServe". It is still going to be released on Monday - and it will have some new stuff. Talking of new stuff, today brings some. Most of it is that "behind the scenes" work I've been doing a lot of - more cleaning of the code and making it slightly better. "!voiceme" has now been totally removed - so, if you want a voiceme plugin, get the one by Eno_ - which is on the Downloads page.

The only other thing that's been done today is something which I call "config versioning" - which sounds more impressive than it is. It just keeps a version number in the config (this number is the version number from the last version of TuxServe used), so it can check if you have upgraded TuxServe - and point out to you that you need to check the config for new settings you have to change. It's not really that useful, I know - but it's just a helpful reminder. It will go away when you save the config - until you upgrade TuxServe again, or change the version number stored in the config via text editor.

Until next update; enjoy.


Aypok...


2004-10-25 :: Voice, Voice, SCREAM!

Just a quick update - not about TuxServe, though. A seperate "!voiceme" plugin is now available. This wasn't written by me, and nor is it the voiceme code from TuxServe - it is written by the amazing Eno_ (so you know it'll work fine!). This has been tested and I can assure you that it works. You can get it HERE. Instructions are included in the "README" file. So, many thanks (once again) to Eno_ :)

I did actually write my own voiceme plugin, based off of the TuxServe code - but largly rewritten. I gave up when it started crashing at totally random points for no reason. The plugin linked to above is better than the one I'd written, anyway.

No updates on TuxServe itself for a few days - I've been very busy writing and recording a load of new songs. Go HERE if you like punk/metal stuff, and check it out.

Enjoy.


Aypok...


2004-10-23 :: A Greenasaurous Ate My Car.

Todays update features yesterdays stuff, as I was too tired to write up what went on yesterday. I'll just get straight to it:

If a DCC chat session (in response to someone typing your trigger) isn't accepted or established within 120 seconds (2 minutes), TuxServe will assume that the user cannot connect and remove them from the list of users. 120 seconds also seems to be the same amount of time XChat gives for a DCC chat session to be established. A related note; if you cannot establish a DCC chat session with them, but they start a DCC chat session with you (before the 120 seconds timeout), it will still work as if you connected to them as normal. But, if you can't DCC chat with people, chances are that you can't DCC send, either - meaning that it is pretty useless :)

Whilst on the subject of DCC stuff: if a DCC send stalls, or fails to connect (times out), TuxServe will class these as failed sends and move to the next one. And whilst on the subject of sends: when a user types "sends" into the fserve window, they will now only see the filename of the sends - not the full path, as before. Also, if you use the command "/ts csaq show_sends" it will look slightly better - and you get to see the full path - the same as with the queues. Another thing about listing the sends: it will also list sends which are connecting or waiting, as well as those which are already connected and sending.

The user commands "clr_queue #" and "clr_queues" will either confirm that the queues have been cleared, or report an error to the user. I noticed last night that users using the "clr_queue #" command can clear ANY queue in the queues list - not just their own. That's now been fixed - it checks if the queue is theirs before clearing it. If it is theirs, it clears it and lets the user know that it was successful - if not, it will let the user know that this is not their queue, and that they can't clear it.

The users can now use "cd ../" as well as "cd ..". If you included the "/" on the end of the "cd .." command, it wouldn't work - it would say that it doesn't exist. That's been fixed. When adding or changing a "trigger_dir", TuxServe will add a "/" to the end of the path if you don't include one. The warning about using long unsigned ints has been fixed, too.

Now for the biggest change/update - something called "show_info". I can't take credit for this idea; Eno_ came up with this great idea. Basically, "show_info" will display information, written by you, about each dir. This may take some explaining, so bare with me :) Say, for example, one of the dirs accessable via your fserve contained a collection of rar archives which are all passworded - how do you let the user know of the password? Put in a text file "password_in_here.txt"? Wait until they ask you for the password? Or use "show_info". When a user enters a directory on your fserve, TuxServe will (if show_info is enabled) check for an "info file" for that directory - if it finds one, it will display the text in the file just above the file listing of the directory. This way, you can include specific notes/info about each directory that will automatically be displayed to the user. So, if we think back to our previous example of a dir full of rars, you can put some text in the info file - something like this:

All the rar archives in this directory are password protected - the password is "monkeys"

Useful, huh? :) Now, the "tricky" part - how to actually use it. Well, first of all, you need to enable "show_info" with "/ts config set show_info 1" - it's enabled by default. Secondly, you need to create the actual info files. These are just plain text files, so open up your fave text editor (pico!) and type out the information that you want the users to see everytime they enter a specific dir. Once you've done that, you need to save it with a specific name: ".TuxServe_info-<dir_name>". <dir_name> must be the name of the directory it is to be placed in (and be careful - it's case sensitive!). If you wanted this info to show whenever a user entered a dir called "lots_of_rars", you'd save the info file as ".TuxServe_info-lots_of_rars". This file must also be in the actual dir which the user is to enter - eg, ".TuxServe_info-lots_of_rars" must be in the dir called "lots_of_rars".

Sorry if that seems a little confusing - I'll write up full docs for it later - after it has been fully implimented. The above stuff works fine - but there is also going to be support for info on each file, too. Please remember to keep the info files reasonably short - I doubt that the users would be too pleased about reading though pages of text. Also, you can include mIRC colour/formatting codes in this text (not XChat type codes, though) - but the default colour is the second colour set in the config (which is blue - or code 12, but default).

It's now been exactly 4 weeks since the latest release, and I'm pleased to say that there have been 137 downloads so far :) The next version, v0.0.11alpha, should be released in 9 days; the 1st of November. All of these changes and updates that you've read about can be found on the CVS repository. If you need any help, or want some stuff (like "show_info") explained, feel free to drop by on IRC.

Update: I forgot to mention the "!voiceme" feature. I'm considering removing it from TuxServe, and either making it a seperate plugin - or just abandoning it altogether. I don't really like the idea of sticking too much stuff into 1 app. TuxServe is an fserve, and a "!voiceme" feature isn't needed. If you want this feature kept in, contact me - if I get enough responses (about 10), I'll keep it in. Otherwise, it's gone.

Anyhoo, that's about it for today - even more stuff to come tomorrow, hopefully. Enjoy.


Aypok...


2004-10-21 :: Phonetical Brick-Praising Day.

Hello there! It is I - Aypok. It's been a while since any news and code updates, so I decided to change all of that. Yep, that's right - there's new code! Yay, and such. And it's all on the CVS.

Today's coding expedition turned up mainly behind the scenes work, but there is some new stuff for you to actually use. The main update was the removal of the "fileio" source files. It mainly contained the old queuing functions that aren't used anymore. The remaining functions were moved to other files - including the new "sends" source files, which contain functions to do with DCC sends and the sends file. Who'd have guessed?

It's now compatable with GCC 2! So, those of you without GCC 3 can enjoy TuxServe. It's been tested and it works fine. Party time!

Now, onto the goodies that you can use. There's not actually much in the way of goodies which you can use today (I was working on something pretty cool - but didn't finish it). You can now ban users by host. So, for example, you can ban "user@isp.com". If anyone trying to log into your fserve matches that, they will be denied access. Also, you can do a broader ban by banning addresses - for example, "some-isp.com". This will deny access to anyone who has the address "some-isp.com" - no matter what their username. This would ban users like "whoever@some-isp.com", "someone_else@some-isp.com", etc - it's like banning "*@some-isp.com". Please note that wildcards are not yet implimented in the bans system. I am, however, planning to add them in the coming days.

If you need to check who's banned, there's now a command to do that:

/ts show_bans

That will list all the current bans into the currently focussed window - only on your screen - no one else will see them. Also, there is no way for the users to list the bans.

Oh yeah - I've rewriten the "voiceme" detection bit. It's now a lot better and easier to follow (if you read the code). It also supports FTP adverts. Here's how it works: it checks each notice it receives for several things - it must contain; "Fserve Active" or "Fserve Online", and both "Trigger" and "Sends" (case sensitive). It should also still detect UPP file servers (which use "File Server" or something - I forget). For FTP adverts, it needs "FTP Active" or "FTP Online", and "Address". If you know of any fserves or other voice-ables that this doesn't voice, please let me know.

That's about all for today - more coming tomorrow. Enjoy.


Aypok...


2004-10-13 :: Flollop Eraser.

A certain "abhilash" helped me with another new bug - one which has now been fixed. When you set just 1 channel to serve in, TuxServe would save it as "all", but still show the correct channel name in the config list. There were a few other bugs in that area, which I discovered whilst fixing this - but they all seem to be fixed. As far as I can see, at least... I've been testing it for about 10 mins, with different settings, and it all seems to work fine. If there are any remaining problems, let me know.

A new binary is HERE (for those of you with pre-v3 GCC) and the source is on the CVS repository. The site has also had a few fixes - updating help and FAQ, etc.

Until the next update; enjoy.


Aypok...


2004-10-11 :: But Who Won The Square Cat?!

I finally got around to updating various bits of the site and the help stuff. I think there are still a few bits I've forgotten about... But whatever - I'll take care of them when I find them.

Also in the news today: it has been confirmed that TuxServe won't compile on GCC versions below 3. So, for now, either; 1) upgrade to a newer version of GCC, 2) click HERE to download a binary of the latest CVS build (extract the latest TuxServe package as usual, and stick this binary in the dir - then use "make install" etc). Or, 27) Wait until the next release - it should be fixed by then.

I think I'll leave off getting back to work on TuxServe for another week - I'm playing around with SDL at the moment (for non-TuxServe related stuff) and there is nothing majorly wrong with this version. Anyhoo, enjoy!


Aypok...


2004-10-10 :: Games And Presents For All!

This is the game I mentioned in the last update. It's basically a version of "Connect4"(tm) for XChat. It can be pretty fun - if you like that sort of thing. I'm not going to go into it too much here, as it's all explained on it's own page; HERE.

That's pretty much it. If you want to see it in action before installing it, drop in to #oldskool on irc.odinnet.ca (during normal GMT hours), and have a game with some of us :)

If there are any problems (with Join-up4 or TuxServe), let me know in the usual way. Enjoy...


Aypok...


2004-10-06 :: Jimmy Has The Bananas!

Today brings more fixes, yes indeedy. Thanks to "[enubis]" for reporting them. The problem is to do with public triggers - or rather, the fact that they ignored the trigger_levels and let banned users in. This has now been fixed and is living a quiet life on the warm and binary-filled CVS servers.

If you get an error, when compiling TuxServe, along the lines of "warning: array `timer_idle' assumed to have one element", then you'll be pleased to know that this too has been fixed, and is living in the same house as the above fix.

Apart from that, not much has happened on the TuxServe front - I'm still on my coding break :) Well, not much of a "coding break" - I got slightly bored and wrote a little game plugin for XChat (more news soon). But I'll be back to TuxServe coding at the start of the next week.

Enjoy.


Aypok...


2004-09-27 :: Yeah! Yay! Yo! And Other Baconated Phones!

It's been 2 days since the release of TuxServe, and all seems to be going reasonably well, I'm pleased to say. One bug was reported - users can still log into your fserve (with CTCP triggers) even if you set TuxServe to be off with "/ts off". That's be sorted and is on the CVS.

Over the passed week or so, I've been trying to help a certain user with a very weird problem. Unfortunatly, it went unsolved :( I'm writing about it here to ask if anyone else has had a similar problem (explained in a sec). The problem was that TuxServe kept complaining about the "ban_list" and "queues" files, saying that it couldn't load them. Yet it could find and use the "config.cfg" file with no problems. We checked that the files were where they should be (and if it can find the config file, then why not the others?), permissions, and all of that sort of thing - to no avail. He was also a user of Mac OS X. I'm not sure if that has anything to do with it, but the path that XChat on the Mac gives TuxServe is different to what it gives it on Linux. If you've had this problem, then please let me know - I want to see if it's a one off problem, or wide spread.

Also, if any of you use Mac OS X, please let me know if the new Makefile works decently. If you find any problems with it, I'd be happy to fix them. Plus, I have no Mac OS X system to test it on, so I rely on you people to inform me of where I go wrong in that department.

As you may know, I've decided to release a new version of TuxServe on the first monday of every month. But, as this latest release was technically 1 week early, I will not release the next version (v0.0.11alpha) until November. This gives me time to finish off my rebuilding of the queues and sends system.

While I'm on the subject of the next release, I'll explain the versioning system. "Why use 0.10, 0.11, 0.12, etc? Why not use 1.0, 1.1, 1.2, etc, and roll over when you hit 0.9?" The revisions of each release are only "minor", in the sense that there are no major changes to the code as a whole - it's basically the same thing, but with additions. Hence I only change the minor version number. Although, when I change the queues system over to the new one, that could be classed as major enough for an update of the middle digit. It's not really important. It's not important at all, actually. I just felt like rambling on about it for a while. I'll just change the numbers when I feel like it :)

I also get a week or 2 of rest from coding. So, for the next 2 weeks, I'm going to be doing other stuff instead of coding. If bugs get reported they'll be handled and fixed as I get them, but no major work will be done for the next 2 weeks.

Here's something you haven't heard about for a long time - the unstable series of TuxServe, or "version 0.1.0". This has been left on the side of the road to rot for a few weeks while I have been taking care of the main releases. Although I will be picking this up again in a while.

If any parts of the site seem broken, please let me know - I accidentally screwed up some of the site. I'll probably be making more "behind the scenes" changes to the site in the coming 2 weeks. For now - enjoy.


Aypok...


2004-09-25 :: TuxServe-0.0.10alpha Released!

It's here! Yay, and such. So, hover on over to the "Downloads" section and downloadinate it. Your chicken will thank you for it. "What's new?" was pretty much covered last night, but there are some last minute updates: You cannot set the "advert_timer" lower than 10 seconds - this is so you don't spam the channel if you accidentally set it to "1". If you've set TuxServe to publically display the advert in the channels, it will display the advert once as soon as it is loaded/switched on. It will then wait "advert_timer" seconds before displaying it again. Thanks to "l00k" for this idea.

Installing this version will be slightly different than before, for some. If you use Linux or some other system that keeps XChat settings/info in "~/.xchat2/", then you install it exactly the same way as before. But Mac OS X users have to do it slightly differently. To compile TuxServe on Mac OS X, you still use the command "make xca", but to install it you use "make install_xca" - this will install TuxServe into "~/applications/x-chat_aqua/plugins/" instead of "~/.xchat2/". There's also a "make upgrade_xca" for those of you whom are upgrading from a previous version of TuxServe.

When compiling this version, GCC (or whatever compiler you use) will display a warning - something along the lines of:

fileio.c: In function `show_queues':
fileio.c:974: warning: unsigned int format, long unsigned int arg (arg 8)

Ignore this warning, as it doesn't affect the working of TuxServe at all. It's an odd thing to get, seeing as how I do the same thing twice, several lines up - yet GCC doesn't seem to care about those... Oh well.

If you get errors with the "ban_list" file whilst running TuxServe, open up a text editor and save an empty file into the TuxServe dir (see above for where it's installed to) called "ban_list". That should fix it. If not, come and find me on IRC. I've only had one report of such a problem, so I'm not sure if it's a one-off, or widespread.

This version, as far as I can see, is relativly bug free. It also seems to be pretty stable (no crashing, or going into infinite loops). All rogue lemur factions that were living in the code have been moved away to some nice lemur-homes. And there are no vast security holes in it! None that I know of, at least...

I think that about covers it. For more help, check out the upgrading help page, and the FAQ. Or come and find me on IRC - I don't mind you people coming and asking for help - so feel free to do so :) Also, please report any bugs you find in this version - no matter how small! Some of this site (mainly the help) is outdated, so I'll be fixing that later today.

Finally, I'd like to thank Moobert for his countless pain-filled hours of help and testing of this version (and previous versions) - without his help this would have taken a hell of a lot longer. So, until next time - enjoy.


Aypok...


2004-09-24 :: W00tage Burgers, With Large Fries.

So, it's been 5 days since I originally planned to release this new version, and it's still not released. But, listen carefully little ones, for there is some good news. I seem to have finally fixed the bugs. /me waits for everyone to stop cheering. Calm down, calm down. "When, in the name of the Holy Frogs of Blorg, is it going to be released?!" If all goes well, and it looks like it will, tomorrow - about midday. Or early afternoon. It still needs to be cleaned, documented, packaged, checked, and baked. Maybe not baked, but you get the idea. So, check back tomorrow - early afternoon (GMT time), and you should be able to get your hairy, sausage eating, paws on it.

"What took so long?" As I've mentioned several times in the past, the queues system - although working - is badly written and plain crappy. I've been in the process of replacing it for the last few versions, and will continue to do so. Basically, some modifications I made to it broke the fragile system. My main priority for the next version is to overhaul the code for that stuff. I was very tempted to do it today, but that would further delay this release, and possibly kill it. Again.

"What else is new? Apart from the fact that users can't browse my entire system and access all my dirty, dirty secrets." Pausing the fserve actually works now. So, if you pause it and people try to get more files, all requests will be queued - even if you have free sends slots. The queues will then resume when you tell it to resume. Also, when people queue files, the message is a bit nicer: instead of just saying "Queue added.", it will now say "File 'filename' added to queue in position #" (where # is the queue position). Obviously, it's in nice bright colours.

Anyone tried resuming sends and queues from a previous session when TuxServe loads? If so, you'll have noticed that it screws up (goes into an infinite loop of writing to the sends file). This has also been fixed. The stats should now report the sent amount of data correctly - although the top speed is still not sorted out correctly. That's on my todo list.

"Is that it?!" Why you ungrateful little... Uh... I mean "no". There's plenty more of the little fixes and changes - the sort of things you don't think matter - until it breaks. I'm sure I had some other stuff to say... But it's 23:48, and I'm kind of lacking in the sleep department... I'll keep this brief and add the proper stuff tomorrow.

This version has no major bugs that I know of. There's a minor "problem" with the setting of triggers for channels - but that's not a bug - I just documented it wrong. That'll be redocumented. Then fixed to fit the current documentation (which is a better way of doing it). Then redocumented back to how it is now. Later.

Yeah, I'll leave this until tomorrow :) I just wanted to share my jubilation. Until tomorrow; enjoy.


Aypok...


2004-09-21 :: The Giblets Will Attack At Midnight!

Just a quick (I hope) update on how it's all going and when the next release will be out. I've managed to fix the permissions problem. I found what it was (not long after I posted the news) and fixed it - so it should never happen again to anyone. Users can no longer download hidden files off of your system if you've set "show_hidden" to "0" (eg, so they don't see hidden files). Plus various updates and bug fixes, including one to fix some problems whereby it checked the ban file way too often.

But, I was correct about the queues system - I have messed it up a bit. Not majorly - it works to some extent - but not in an overly useful way. I doubt it'll take long to fix this. Although it may delay the release by a day or 2. I refuse to release it until I'm happy that I've fixed the major bugs - otherwise, there's no point in releasing it (as this is mainly planned as a bugfix release). There are a number of other things that need to be done before the release, but I'm confident that they'll be sorted out without problems.

Now, on to the release itself... I was initially hoping for a release of it yesterday, but evidently that never happened. Today and tomorrow were also possibilities, but I've got other stuff planned for all of today, and I doubt i can do all the code for a release by tomorrow... So, to be honest, I'm not sure when it'll be out. It needs be released ASAP, but it also needs to work properly... Conflicting interests. I'm now planning for a Friday (24th) release - that seems like enough time. But again, I make no promises.

Also, if you have any bug reports or feature requests; now is the time to submit them, otherwise you'll have to wait for another version. One last thing: does anyone know of any good (and free!) shell providers that allow outgoing connections to IRC servers? This would make testing a lot easier for me. I know that I could just google - but I'm hoping for some recommendations :)

For now - enjoy.


Aypok...


2004-09-18 :: Blah...

Hey peoples (yeah - apparently there are people who occasionally read this!), some good news, and some not so good news. Oh, and some more good news. Let's do it in that order. First, I seem to have fixed the security problems whereby a user could browse your entire system. How the Hell something like that got by me, I don't know... But it's fixed. I'm pretty sure it's fixed... If it's not, feel free to shout at me.

The bad news. Well, not that bad - it's only temporary - I hope. Any of you remember that I reported about some people getting "No permission on this file" errors on files they had permissions on? Well, I've now got that problem. Yep. One minute it was working fine, the next minute I don't have the right permissions. The root of the trigger dir shows fine, but none of the sub dirs work. I have no idea why :( I know it's something to do with the "stat" command - used to check file attributes and such. I'm hoping to fix this soon - but I've given up for today... My brain's had enough (see below).

Oh yeah, before I forget - I might have messed up the queues system. Then again, I may have not harmed it at all. I was looking over the queues system, making some modifications, and I realised some bits seemed "wrong". Just as I went to test it, the above "permission denied" errors reared their heads, meaning that I can't test it.

Good news: There's now a command to allow leechers/downloaders to check their queues without having to log into your fserve every time! It's the new "TuxServe anti-n00b system". Or not. It sounded like a funny name... Anyhoo, all that the user needs to type is "/ctcp <your nick> my_queues" and it will send them a list of all their currently queued files, along with the queue position of each file. The list is sent in the form of a notices - which will be invisible to you if you have set TuxServe to use raws. There is currently no option to turn this on/off (it's stuck "on" atm), but I'll add one in the future.

Other little things that have been done include; updated the showing of queues (as well as the above CTCP queues thing) - users no longer see the full file path of the files in the queue, just the file name. The person running TuxServe will still see the full path of queued files. When a user tries to change directory (with either the tags or the "cd" command), but fails, it will not relist the dir as it did before. It will now automagically add a space to the end of channel lists if you forget to add one when configuring TuxServe.

Ever had the problem where users log into your fserve, log out, but can't log back in? I've known about this for a long time, but I've never know why it is, or how to fix it. I thought it was a fault in the users system - so I added a function to clear all users. Although this is useful, I don't think it fixed the problem. But I found out what seems to cause it: If a user tries to log in whilst you have a dialog window of theirs open (eg, from a previous DCC chat session), they will not be able to log in. If you change focus to a different window, such as a channel window, they should be able to log in fine. I'm still not sure why this is, but I'm looking into it. Although it's of a lower priority than the above problems.

I was going to write some stuff about a damn maths problem hurting my brain... But my brain hurts too much. If you want to see what I mean, here's a link: http://ircimages.com/popup/ircimages/5/d/5d36a997f2804f19f0abe8df0bcc75df.jpg. My problem with it is not the "what's this hole" bit. My problem is with the fact that the sum of the area of the 4 shapes in the top triangle, do NOT equal the area of the top triangle as a whole. Ignore the bottom triangle - I really don't care about that... Please, if you can explain what the Hell is going on, drop by on IRC - several friends and I are very keen to understand it :)

Finally, tomorrow (the 19th) is national "Talk Like A Pirate Day". The European (UK) site is here: http://www.thomasscott.net/yarr/. I don't know where the US or international sites are - but i'm sure Mr. Google does. Should be fun. Anyhoo, I think that about does it for today. Enjoy, me maties (just practising).


Aypok...


2004-09-17 :: Progress, and Such...

Good news everybody </Farnsworth>, the security thing is fixed. As far as my testing shows, anyway. When the user tries to go up a dir from the root of the trigger_dir (either with the tags or with the "cd" command), they will receive an error informing them that they cannot go up a dir. There was another small problem where the user could get around this by using a command such as "cd ../..", but that's been fixed (I think - it has yet to be fully tested).

So, that's all on the CVS repository, if any of you want to get it right now. A word of warning for those of you who intend to use this CVS code - there's an option called "show_hidden" that is semi-implimented. By default, it will be "0" - MAKE SURE YOU SET IT TO "1"! If you fail to do this, it will not show the hidden files. "Why's that bad?" - the tags system does not account for not-shown hidden files! Making the tags system useless.

"When's it available to us mortals who wish not to use the CVS?" I'm planning to have the new release pack out early next week. I want to say Monday 20th September, but I'm not sure if that gives me enough time... It'll be between the 20th and the 22nd of this month. It will be a full release (version 0.0.10 alpha) as opposed to a bug fix pack. It will also, hopefully, have the hidden files system fixed, much better Mac OS X support (see below), and general fixes in the way you configure it (making it easier). Probably more stuff too - like the banning command, which now works.

Now, about Mac OS X support: It's come to my attention (thanks "QueerMe") that the TuxServe install script in the Makefile installs it in the wrong place for you Mac users. You'll have to forgive me for such errors as I don't have a Mac OS X machine to test on, so I'm relying on help from you users (which has so far been great - thanks!). If I understand what I've been told correctly, TuxServe (and its files) need to be installed in the "~/applications/x-chat_aqua/plugins/" directory. If installed in the default location (eg, where TuxServe currently sticks itself), it will cause errors like not being able to find its config file.

As always, if you need help, find an error/bug in the code/site, have some suggestions/ideas, or want to offer some help - please feel free to contact me and let me know. You can use the forums, the bug tracker, email, or - my favourite - IRC :) See the Contact page for more information.

Enjoy.


Aypok...


2004-09-14 :: Bug Warning!

I've just found a very large bug in TuxServe v0.0.9 (and possibly earlier versions)! If the user browsing you fserve is in the root of the trigger_dir (eg, where they start), THEY CAN GO UP A DIR AND BROWSE YOUR ENTIRE SYSTEM! I found this late last night whilst getting back to work on the code (first time in 2 weeks). I was hoping for this first post in a week to be good news :(

I didn't post info of it last night as I was trying my best to fix it, and planning to then post a fix for it in a short space of time. That didn't happen (my attempts to fix it have thus far broken the entire browsing/getting of files) - so I thought it best to warn you of the danger.

"What do we do?!" I can think of 2 answers for now: 1) Make sure the user you're running XChat/TuxServe under doesn't have permissions on any files/dirs outside of the serving directory. This is probably more trouble than it's worth - bringing me to option 2) Stop using TuxServe until I fix this - I'm spending every minute I have (except now, to warn you) trying to fix this.

You could, if you want, keep running TuxServe as normal - how many users will know of this problem, or try to go up from the root of the trigger_dir? "Not many" is my guess. But, it's better to be safe than sorry.

I apoligize for this rather large screw up and to those of you whom are affected by this. Be sure that there will be a fix shortly. I may just release v0.0.10 early, instead of a bug-fix pack. Keep checking back here to see if there's any progress.

Enjoy?


Aypok...


2004-09-06 :: TuxServe-0.0.9alpha Released!

Yep, just as promised - the latest version of TuxServe. This version has a lot of upgrades and bug fixes over the previous version - so if you're using a previous version, this upgrade is highly recommended. What's new? Loads of stuff :) You'll have to check out the changelog to find that out (or just read the old news on this site).

The new version can be found on the download page, on the left menu. If you're using a previous version of TuxServe, you will want to read the page on updating TuxServe, HERE. This is another new page. It's not because upgrading is hard or particulary complex - it's just there to help the upgrade go a bit more smoothly.

There's also a new page for this site - "Security". This answers some questions relating to any security issues relating to TuxServe, and how to make it more secure.

There is, however, some bad news - TuxServe has not been updated since last Sunday - over a week ago! It was a busy and crazy week - the first half was spent in the real world, with friends and such - but I was determined to spend Friday, Saturday and Sunday coding away to get some cool new stuff done. My main hard drive decided otherwise...

I had been expecting it to give out for a while now, so I bought a replacement drive - but just as i was adding in the new drive, the old drive died. The plan was to add in the new drive, backup the broken drive, and then reinstall Slackware sometime on Monday (today) after I had released TuxServe. But after hours of trying to fix the drive, I gave up and spent the next few days backing up as much data as I could and reinstalling.

Fortunately, no data was lost (well, apart from the system files - like the contents of /usr on my old boot drive) - and all the TuxServe code/data was unaffected (and I had it backed up on the CVS anyhoo). Also, I now have a nice fresh installation of Slackware 10.0 (I'd been using a year old installation of Slackware 9.0 which was very broken - it was actually installed on a completely different computer, then I moved those drives into this one :) ). Not everything is fixed yet (not got any IDEs loaded) - but I'll sort that out and hopefully get back to coding with a day or 2.

Anyhoo, that's enough crap about me :) So, why did I bother saying all of that? The bans still don't fully work. Yep. didn't get a chance to fix them :( Give me a few days.

Don't forget to DOWNLOAD THE NEW VERSION! Enjoy.


Aypok...


2004-09-02 :: The First Century

Party time! TuxServe 0.0.8alpha has now officially (according to the SourceForge stats) been downloaded over 100 times! Okay, okay - that may not be very high compared to some projects, but I'm pleased :) /me does a little dance

And the new next release (0.0.9alpha) is out in 4 more days. /me does a little dance

No updates since Sunday, though - I've been busy with this "real life" lark, but the release will be fine. I'll have some time tomorrow and over the weekend to add and fix more stuff. /me does a little dance

Here's a mildy amusing screenshot that I took a week or 2 ago. It is genuine, and not edited (except the cropping) in anyway. I found it while browsing SourceForge and decided it was odd enough to take a screenie of. So, HERE it is. If you are wondering what's funny - they are in order of project activity :) /me does a little dance

As usual - enjoy.


Aypok...


2004-08-29 :: Run Rabbit Junk

Hey hey peoples - today has been another productive day. Yup. I'll just get straight to the list of new stuff:

/ts kick <nick>

Kicks <nick> from your fserve (they get a message saying they were kicked). All their sends and queues are unaffected by this - it's just used to get them off the fserve. There is currently a bug due to the fact there is no error checking (coming later) - if you try to kick a user that isn't logged in to your fserve, it will display the kick message in the current window, which will probably be a channel. So be sure you get the nick right :)

/ts ban <nick>

/ts unban <nick>

First off: the ban command does NOT work yet. Not sure why, but I'm working on it. The unban command works fine... Banning works fine, just the command to add bans is broken. You can add bans manually to the "ban_list" file in the same dir as the rest of TuxServe (~/.xchat2/tuxserve/) - 1 nick per line. This will ban them from all triggers (CTCP and normal "!style" triggers). They will still be able to see your fserve advert via "!list" (see new options below). Currently, only nicks can be banned (dispite what TuxServe itself says) - but host banning to be added in the coming days.

/ts pause

/ts resume

The former will pause ALL DCC sends from you (by killing them), and sends a message to everyone who is being sent to that their sends will be resumed later. The latter command does exactly that - it resumes from pauses (or a complete shutdown of TuxServe). Please note: these commands are, at the moment, UNTESTED. Well, the pause command worked to some extent - but had bugs. Beware for now.

When starting up TuxServe you will notice that it will say 1 of 3 new things (about queues and sends). This is because when you start it (if a certain option is set - see below) it will automatically check the queues and sends files - if it finds anything in them, it will resume sending them straight away.

The new options/settings are: (the numbers after them are the default settings: "0" is off and "1" is on.)

auto_resume 1

This is for when TuxServe starts (if it's set to "auto_on") or when you switch it on with "/ts on" - it will, if set on, automatically resume all sends/queues it finds in the files. (It'll currently only resume upto 1 queue at first, but should sort itself out when another send finishes).

auto_clear 0

This is sort of the opposite of the above option - this, if enabled, will automatically clear all queues and sends when TuxServe is started/switched on.

inform_banned 1

This is briefly explained when you ban someone from the fserve - if this is on and someone who is banned uses one of your triggers to get into your fserve, TuxServe will message them to let them know that they are banned.

banned_no_list 0

This stops all banned users from seeing your fserve's advert when they type "!list". They don't get a message letting them know you have an fserve that is invisible to them.

Other additions/improvements include: the "cd" command is now fully operational. I've added some more help to the help section. Made the fserve commands the user uses case-insensitive (so they can type "help" or "HELP", or "HeLp") - this includes tags: "1f" is now the same as "1F", etc.

That's quite a bit of stuff :) It's all on the CVS at the moment. I'd like to quickly point out something about the CVS repository: If you checkout the source before about 18:00 GMT, then it will most likely have some fatal errors, as it'll be testing code that I'm working on. After that, I'll have either fixed it or at least made it usable :) You could, of course, wait another 8 days for the next release of TuxServe.

I'm not done with this news update yet... Next is the site. The help sections are soon to get upgraded. First off will probably be the FAQ - a load of new information that will be useful. Secondly will be a guide for end users - eg, people who use your TuxServe fserves to download files - just giving them a quick guide on the differences (and how they affect them) in TuxServe compared to other fserves. I'll also add all these new commands and options to the current help docs.

If anyone has any suggestions for the site (or TuxServe!), please contact me - I'd love to hear them. For now; enjoy.


Aypok...


2004-08-28 :: Little Things

I've spent most of this morning coding on the stable release of TuxServe, adding a few little things, and some slightly bigger things. Most of the things were code tidying and stuff you won't really care about - but there are a few noticable changes (although minor).

The biggest change of all is that the "get" command is now implimented. As you know, users had to use the tags to get stuff from your fserves - not anymore. I know there's no point in having it, as tags are a better system - but it's there just incase any users need it. When "get" is used for the first time, it will give a brief explanation of the tags to the user and inform them of the "help tags" command that explains, at some length, how to use the tags. Whether the user chooses to use the tags or the "get" command, the outcome will be identical, as they both call the same functions in the end :) The "cd" command is still broken - but I'm working on it.

Another of the larger (yet not visible) changes is that I'm using the new queues system - well, most of it. The queues system in the unstable series was written completely from scratch and works flawlessly (as far as I can see) - so I decided to backport it to the current stable series of TuxServe. This way I can scrap the God-awful current queues code and make it much more maintainable. And it should be a lot less error prone.

As quickly mentioned before, the help system for users browsing the fserve has been beefed up a bit. When they type "help" they get a list of 2 new options and a link to this site. The new options are "help tags", which gives an in-depth explanation of the tags system and how to use it, and "help license", which will display the GNU/GPL - the license that TuxServe is released under.

Also, when a user (in the fserve DCC chat window) types "sends", the list of sends (like the list of queues) will show the sizes in relative terms, eg - it will show the size in bytes, kb, mb, or gb, depending on which is approriate.

The rest of the changes are either cleaning some variables or little cosmetic things. Every little helps :) The day's only half over, so I should be able to get some more coding done - off I go...

One more final thing: The "known bugs" list isn't exactly upto date. I'll try and get some time to update it soon.

I'd also like to thank everyone for their support of this project - you guys and gals are great :) Enjoy.


Aypok...

2004-08-26 :: Generic Title

More bug fixes today - timer ones again. If you had a problem where, if you changed a timer the old timer AND the new timer would both run, then the answer is on the CVS. Another bug spotted and reported by "IkaTaii". /me should give out some medals, or something...

Added some more functionality to the unstable version too. I was planning to spend all day doing so, but I accidentally wrote a cool new song, then spent the rest of the day recording it :) If any of you are interested in checking out my music (punk/grunge stuff), check out my other site HERE. The site looks awful, I know - but it does the job... The new song isn't on there at the time of writing, but it should be up by the end of the day - it's called "Nuclear Physics" if you want to look out for it. It's not great, but meh - I like it. The best track is "Hunters End".

Righty dokey, I think that's all for now. Plus I need food and stuff. Enjoy...


Aypok...

2004-08-25 :: Bug Fixes All Round

Another bug, once again it was in the config system, was spotted by "IkaTaii" and has been fixed. Anyhoo, the update has been in the CVS repository since last night. This bugs stops you from setting the serving channels (and the voiceme_channels) to "all".

You may be wondering where the bugfix pack is - well, there isn't one. And there won't be. That's part of the reason I delayed updating the site until today (that, and I was very tired).

The reason for the lack of a bugfix pack is that at this stage after a release, pretty much everything has been updated or modified - so to release a bugfix pack would be the same as a normal release pack. So, I've decided to save it for the next main release, which will be version 0.0.9alpha, and will be out on 2004-09-06. All updates are, as always, on the CVS - so if you're desperate for the fixes, please get the CVS code.

Enjoy.


Aypok...

2004-08-22 :: Posting Like Crazy

A news update every day, eh? Yep, there seems to be quite a bit of work going on - and most of it is down to you TuxServe users for your suggestions, ideas, bug reports, and patches. Keep them coming! :)

There's now an option to automagically close fserve DCC chat dialogs when a user types "quit" or "exit" in them. By default this is off (eg, leave the windows open), but it's there for those of you who want it. You'll need the latest CVS code for this feature. Here's how to use it:

/ts config set auto_close 1

Or, replace the "1" with a "0" to turn it off again. The idea and original patch for this was by "hagar81", not me - so thanks once again to him :)

Another suggestion of his was to include the usage of "raws" instead of sending all commands to XChat for it to carry out. This is now implimented. All "notices" from TuxServe (message like "too many users", "you need to be voiced to access this", and the sending of the advert when someone types "!list") are now, by default, sent as raws. This means that when TuxServe responds to a "!list" by sending your advert, you won't see your advert sent. Nor will you see any of the notices it sends. All PRIVMSGs (eg, when you forcefully display your advert in a channel, or when the timer makes it show) are still visable to you. If you want, you can turn off the usage of raws for sending notices and let XChat handle them (meaning that you see them). This command will disable using raws:

/ts config set use_raws 0

To enable the usage of raws again, replace the "0" with a "1". The only real difference between using raws or not is that you don't see the messages.

If this rate of change/development keeps up, it'll be time for a new release (of the stable series) within a week or 2 :)

Enjoy.


Aypok...

2004-08-21 :: Semi-random Posting

Got stuck on some coding bug and decided to take a break to do other stuff - like post this. There is some "real" news to report, though:

The Makefile now includes a "make upgrade" option. The basic "make install" option over writes all previous installations of TuxServe - including the config and queues files. This new option only upgrades the TuxServe binary and keeps your settings/queues intact. This may be useful when, for example, there's a bug fix meaning you need to recompile TuxServe - just do a "make upgrade" and it will give you the fixed TuxServe, but keep your settings.

There's been a fair bit of work on the unstable side of things, too. I've started work on the actual XChat plugin part of it. I have hit some minor coding problems - but nothing that can't be fixed :) I may have a version that "works" ready within a week or 2 - but that's not a promise :)

There's also a "Features" page for TuxServe HERE. This is the only current link to the page - I'll add proper links later.

Enjoy.


Aypok...

2004-08-20 :: The Return Of Eno_

I've re-GIMPed all the button graphics and made a few more changes to the layout of certain pages. There are 2 new buttons: "Bugs" - which takes you directly to the bugs page - and: "Links" - a list of other sites that I think are worth mentioning. Note, that the links pages doesn't quite exist yet.

You know that problem where TuxServe spews everything into the current window/channel if you don't set the XChat option "set gui_auto_open_dialog on"? One "hagar81" suggested a simple solution to that problem, and I've now implimented it - so thanks go to him for that. Basically, when a DCC chat session is initiated, TuxServe will "query" the person trying to connect - thusly making sure there is an open window to send the text to. This is on the CVS, but if it's requested, i'll make it into a bugfix pack :)

Enjoy.


Aypok...

2004-08-19 :: Site Updates

As you can plainly see (unless you use lynx, or have images disabled), I've changed some of the site's graphics. The main tables now have borders! Okay, okay - they may not be the best looking, but they were much better than the normal table borders. These were quickly done in The GIMP this morning, and may only be temporary - so don't be surprised if the borders have changed by your next visit :)

I also decided to make proper use of PHP (well, for something like this you have to) - so the entire layout and look of the site can be changed in a few lines of code :) I'm very pleased with all that - much more so than with the graphics. Not that you get to see how cool it is :).

Now, back to TuxServe... No new bug fixes (no new bugs reported), but the unstable version is coming along very nicely. I managed to get a large amount of stuff done yesterday.

Anyhoo - my hands are hurting from typing all morning. Enjoy.


Aypok...

2004-08-16 :: Busy, Busy, Buggy

Today brings more bug fixes. Thanks again to IkaTaii and Whiffletree for spotting the bugs. The first bug causes channel settings (when used with a command like "/ts config set channels #serve-here #w00t #etc") to either not be set, or to set the "voiceme_channels" to this, instead. The second bug is to do with setting the triggers/trigger_dirs - when you set them, they'd do one of several things: be set correctly, but unset the following setting; ignore what number you gave it, and add itself to the end; or not set anything at all.

These bugs have now been fixed - go to HERE to get them.

A quick note about TuxServe bug fixes in general - you only need to apply the latest bug fix, not all available fixes - the latest fix will contain ALL previous fixes for that version.

Until next time; Enjoy.


Aypok...

2004-08-15 :: More Stuff

There's now support for Mac OS X - you don't have to edit the Makefile yourself. Thanks to Mike Goggin for this! The new Makefile is on the CVS if you want it (instructions on using the CVS repository and compiling on Mac OS X are in the "Help -> Getting Started && Installing" section).

The CVS has been split into 2 modules now - the stable series is still as it was and the instructions on how to get it remain the same. The new module "unstable" is for the unstable version that I'm currently working on. Keeping them in the same repository was starting to get annoying. Instructions on how to download the CVS of the unstable version will be coming shortly. There has been some more work and improvements on the unstable version - but it's still nowhere near ready :)

The final bit of news for today is about this site itself - there've been several new pages added and I'm working on repairing the code behind the pages (I hacked them up pretty quick) - so you may notice a few changes. Nothing major, though. The new pages are:

Bugs - A list of all bug fixes available.
Thanks - A list of all the people who've helped me out with TuxServe.

I swear there was another one... Meh, oh well :) Enjoy.


Aypok...

2004-08-13 :: Progress And The Future

First off, a bug fix - this fixes the problem with the advert timer not being set to what you told it (eg, if you set it to 3600 seconds, it'll set itself to 16). Click HERE for the bug fix and how to apply (easy enough) - thanks to "skamp" for reporting this (BTW, sorry I missed you on IRC - I was asleep. But I got your message :) ).

A quick note about the 0.0.x series of TuxServe (or "stable") - this is still supported, so please keep letting me know about bugs - I will try my best to fix them. I will keep doing so until 0.x.0 ("unstable") is usable and replaces 0.0.x - which won't be for a while (see below).

I've been adding to TuxServe 0.1.0 over the past few days and updating the CVS as I go - there are no plans to stop doing this, either :) Progress over the last few days has been a bit short due to me finding another Half-Life mod, called Sven-Coop (which is great fun)...

About the CVS: Currently, the CVS contains the latest stable release (0.0.8alpha) with all bugfixes and improvements, etc. It also contains the latest 0.1.0 code. So, when you do "make" - which one does it compile? The stable one: 0.0.8alpha. When you checkout the CVS code, you'll notice some new file, ones that don't seem to be used - they're some new ones (parts of the new version are rewritten from scratch). How do you compile the unstable version? "make d". Please note that the unstable version does not work yet (by that, I mean: is not usable) - that is why the stable series will still be maintained. I'll probably even release another stable pack: 0.0.9alpha :) But not for a while.

So... This "big plan" for the new TuxServe - what is it? Well, here goes: Basically, TuxServe will be in 2 parts - the "back end" (which does all the file handling - reading/writing the config/queues/sends files, etc) and the "front end" (the actual plugin for XChat, which handles the displaying of adverts, getting user input, activating DCC sends/chats, etc). The back end part will be a daemon (a process running in the background) and will communicate with the front end via sockets.

This may seem a bit over the top to some people, but I think it's worth it - plus XChat plugins are annoying to debug - this way makes as little of the code XChat specific. This brings around another benefit: TuxServe will not be limited to being just an XChat fserve. All that will be needed is for the XChat front end to be converted to something like an Irssi plugin (as a lot of the work is done by the back end, this won't be too hard) and it will work the same. Technically, it's not just limited to IRC - someone could write a plugin for GAIM, which uses the AIM/Jabber/MSN-Messenger protocols to send files, but uses the TuxServe back end. Of course, that's theoretical, as I will just be working on the XChat part. If, however, anyone else wants to write a front end for TuxServe for another app - please do. I'll try and help as much as possible with it. Or, you could fork the project and take it your own way (thusly, no interference from me) - it's upto you :)

This also brings us to the point which most people want - a GUI based config tool. Yes, yes - I know that the current way of setting the config is a pain - even buggy - that's why a GUI one will be in the works. This will work in the same way as a plugin (connecting to the back end via a socket) - and be GTK based. If you want to write your own, or make one using another tool kit (QT, etc) - as above, feel free. However, I currently know hardly anything about writting GTK apps - so that's quite far off. Plus, the back end isn't finished yet.

As I hinted to above - the new version isn't finished yet, so writting a config editor or new front end for it at this stage would be a bit hard (maybe). I'll be posting news of how it's progressing as I go along, and if anyone is interested - some guides on how to write a front end for TuxServe (eg, how to communicate with the back end).

Quite a long news update today... But it's not over yet. Some people are having trouble connecting to Odinnet (the IRC network that I live on) - one of the servers is down for repairs. If irc.odinnet.ca doesn't work, try: edmonton.odinnet.ca (or tampa.odinnet.ca).

Thanks again to everyone who contacted me about bugs/problems with TuxServe - I really do appreciate you taking the time. And if I'm not around on IRC, either wait a few hours, or just PM me the problem/info. I hope I haven't forgotten anything... Oh well: Enjoy.


Aypok...

2004-08-03 :: TuxServe 0.0.8alpha Released!

And about damn time :)

So, what's new? Well, the queues work, and everything is now brightly coloured :) For detailed, uh, details, check the "changelog". It's mainly little improvements over previous versions - nothing major.

So, if you've been waiting out for a working version of TuxServe before trying it, now is the time. Also, if you do use it and find a bug, then please let me know - either via the SourceForge bug tracker, or contact me (see the contact page) - I'd really appreciate it. Feel free to send me feature requests or ideas, too.

For now, i'm off to work on the new version. I'll be making a new page over the coming weeks detailing the new version - how it'll work, and how you guys can help :)

Get the new release - and enjoy :)


Aypok...

2004-08-01 :: Long Time Coming

What is this?! A new news item on the TuxServe site? It'd seem so. So, what's going on? Good things!

Yup, good things indeed. Development on the next version (v0.1.0) is coming along and will be appearing in the CVS repository in a few days. But in the mean time, I have a treat - and what a treat it is...

It's a working, yes; WORKING version of TuxServe! Yep. Even the queues work!. How's about that? :) It's the long overdue "TuxServe 0.0.8alpha" version. Its release was hindered by the queuing system that I cobbled together.

"Why now?" you ask. When asking for help with a certain aspect of TuxServe 0.1.0, I suddenly figured out the problem with previous versions :) (I'd like to thank Eno_ at this point for all his help :) ) I hacked the fixes into it, and sure enough - it worked! I took the oppertunity to add another few bug fixes into it, too.

"So... Where is it?" I hear. Wait a day or 2. I have to do a bit more testing and add a few more fixes, then it'll be out. After that, I'll be spending my time on 0.1.0 - which will have a lot of code written from scratch. So, it won't end up looking like the current "fileio.c" file *cringe* - that's an awful mess... I've already begun work on a much nicer replacement :)

That's all for now - check back in a few days. Enjoy.


Aypok...

2004-06-04 :: Finally Back Online!

Hey! Sorry for the lack of... Well, everything... But I couldn't help it - I was offline. During a large, countrywide, storm - just over a month ago - lightening struck the phone line just up the road from here. This sent huge surges down the phones line to the houses in this area, and blew up modems which were plugged in at the time. Mine was one of them blown up - and 2 of my friends had this problem, too.

However, that wasn't the worst of it. My modem, for some unknown reason, sent this power surge back through the USB connector and into the computer. This blew up the PSU of my gateway box. It also fried the motherboard and CPU... I had a spare board and CPU, but alas - the USB controller on that system was trashed.

I spent my time over the next few weeks browsing eBay from a friend's computer (thanks Gazmor!), and managed to get a new modem and computer (I couldn't be bothered to wait for the right components to fix my broken one - I just bought a complete P166 system). And that is the story of how I lost, and then regained, my 'net access.

There is yet more bad news: During this time, no (well, very little) progress was made on TuxServe. This is because I need to be online to test the new code for TuxServe. There's no point in me writing a load of new code on the top of some stuff that doesn't work.

That's all sorted now! So, back to coding I shall go. Not only that, but I've had some great new ideas not only for features within TuxServe, but for the workings of the system as a whole - so expect MAJOR changes in the coming weeks :) These new ideas aren't finalised yet, so I cannot give you details.

I appologise for the halt in development, but as I explained earlier - I couldn't do much about it.

For now; enjoy.


Aypok...

2004-04-25 :: More Goodies

I managed to get a few more things added and fixed yesterday. Nothing major, but every little helps :)

The new additions to the queue system seem to have back-fired a bit. Now, instead of not removing any queues, it removes them all. I think I may have fixed this problem, but the fix is yet to be tested.

There are several new commands that have been added. I haven't documented them yet, but I'll be sure to do it prior to the next release. The new commands are:

/ts csaq add_queue <nick> <file>


<nick> is the nick of the person you want to queue the file for, and <file> is the filename (and FULL path) of the file to be queued for the user. This adds the file to the end of the queues list. It also messages them to let them know it has been queued for them.

/ts csaq queued <#>


This sets the number of files queued. It's only really of use for debugging and fixing errors. It doesn't remove nor add files to the queue - it just changes the variable that stores the number of queues.

/ts csaq sending <#>


The same as above, but for the number of files currently sending.

When listing the queues (either from the command "/ts csaq show_queues", or the user typing "queues" in the fserve window) it should now show the size of the queued files - and it should not just show bytes, it should show kb, mb, or gb - depending on the files size. The same goes for the sends list - except that currently only shows the files size in bytes.

More coming up - but for now; enjoy :)


Aypok...

20-04-04 :: The Code Floweth

The first CVS update for a few weeks happened yesterday, as I got started on TuxServe v0.0.8alpha. There are a few new things and bugs fixes included. Such as: a completely new function to handle queue removal (old versions didn't do this correctly) which should work. Although, it's thus far untested :)

I've also tried to make it look nicer - yup, I made all the text that TuxServe displays coloured (which can be changed via the config changing stuff). Most of the colours look good (in that they make stuff stand out and easier to read), but the colours the users see when they log into the fserve are also unseen. I'll be testing them on some unsuspecting victims within the next hour, so I can see how they look and make changes accordingly.

Other minor things have been updated - like the problems you get if you miss the "/" off of the end of the trigger_dirs, and such. Not bad for a days work. If you want to check it out, it's on the CVS right now. I'll probably adding a bit more code later today, too :)

Enjoy.


Aypok...

19-04-04 :: Back to Business

It seems that my plan of getting back to work last week didn't work out, hehe. But I've decided that it's about time to do some more coding. So I shall be updating TuxServe, the CVS, and probably this site, all this week.

I've also become aware of a potential "problem" with XChat. I put that in speech marks, as it's not an actual problem or bug, but it can seem that way. There is an option in XChat called "gui_auto_open_dialog", and this NEEDS to be set on for TuxServe to work. If you're not sure how to do this, type (in the normal text entry area):

/set gui_auto_open_dialog on


It should say something like: "gui_auto_open_dialog set to: 1".

"So, what is this 'problem'?" you may well be asking. Well, if the aforementioned option is set to off, XChat will not open a new dialog window for new private messages nor DCC chat sessions. So, when TuxServe starts a DCC chat session with someone wanting to browse your fserve, XChat does not create a new dialog window - and this confuses TuxServe. TuxServe will then display everything it's supposed to send to the user, in the current window - which will probably be a normal channel - as happened to me :)

I'm going to hopefully sort this out soon by making TuxServe create a dialog window if none exist - but for now, this will have to surfice. This info (and a few others things) have been added to the FAQ.

Anyhoo, I'm off to code TuxServe. I haven't even looked at any code for 3 weeks... I hope I remember C :D Enjoy.


Aypok...

04-04-04 :: TuxServe-0.0.7alpha Released!

Yep, finally. I was planning to release it this weekend, since about a week ago - and now I have! :) Not much has changed since the last update, thought - I've been busy all week with other business. This same "other business" will also be taking up the coming week, but I will see to it that I get back coding after next weekend. So, in the mean time - you get this :)

"Does this version actually work?" Yes! And no... Users can go in (from the advert - which all works), browse the dirs you let them, and queue up files - then they send. And it works! I've seen it work too :) There are a few minor problems, however. The queues files does not always get cleared after a queue is sent - this is my main priority, and will be fixed before all other bugs. So I'm on that. Also, the "autoon" feature is broken - so whenever you load TuxServe, you must manually switch it on with the command "/ts on" (see the help section of this site for more info). There are other bugs too - see the "todo" file in the release pack (available on the downloads page).

Now would be a good time for testers to start playing with it and giving me bug reports and suggestions. You're the people who are going to be using it, so I want to know what you want it to do :) Any help I get from you guys'n'gals is greatly appreciated.

Anyhoo, i'm off (I'm busy atm) - enjoy.


Aypok...

29-03-04 :: News Update.

I was planning to write this update over the weekend, but I got sidetracked in some stuff, so meh. But anyhoo, there's some good news to report :)

Some of you (who am I kidding? No one but me uses TuxServe :D) may have noticed that sometimes TuxServe doesn't respond to "!list", or to "!voiceme", etc. This problem has now been fixed. /me will now explain the problem just for the hell of it. The problem was that when you join a channel, the case of letters used in the channel name matters - eg, "#oldskool" is not quite the same as "#OLDSKOOL". The IRC servers know what you mean and they don't care about the case, so you'll still get to the same channel, but when your IRC client sends a PRIVMSG (eg, text to a channel, or a private message to someone), it sends the name of the channel as you typed it (with the same case). This may not seem like a big deal, and to be honest - it isn't, but it annoyed the hell out of me until I finally found the problem. The problem was that I was using standard case sensitive string comparing functions, so when TuxServe checked its list of serving channels and read "#oldskool", then it read the "!list" from someone in the channel which was from "#Oldskool", it thought they were different channels, and thus - didn't respond.

But that's all fixed now thanks to some case in-sensitive string compairing functions. I also fixed other case sensitive issues, such as "!list" not being the same as "!List". Aren't you glad you read all of that? :)

I was planning to have the beta version of TuxServe out by the past weekend, but I decided against it (not just because I was busy). All the major parts of TuxServe now seem to be working properly (yay!), but I'm doing a bit of "field-testing" first, to try to find and fix as many bugs as possible before releasing it as a beta. You can still get the latest version on the CVS, with all the bug fixes and stuff.

But for now; enjoy.


Aypok...

22-03-04 :: Site Update.

I've just added a LOT of helpful documentation on TuxServe to this site!

FAQ - I added this a few days back, but I just updated it.
Installing - This is the help doc from before.
Configuring - This is a large document detailing all the TuxServe settings and how to change them.
Using - A list of all commands TuxServe accepts.

You can browse these by clicking the "Help" button down the left menu. The FAQ can be found from the "FAQ" button (duh).

The configuring document is not quite complete - it's missing the info on how to set up the advert settings. This will be added shortly (tomorrow or the next day). The rest of it is complete. If you have any questions you think I should add to the FAQ, or you have some questions that these docs don't answer, then please contact me :)

Enjoy.


Aypok...

21-03-04 :: It's Coming Together.

Over the passed few days, as well as coding TuxServe, I've been writing informative guides on how to use and configure TuxServe. These are not quite finished yet, due to the amount of things to document, but they will be available shortly.

Another thing that is coming together, is TuxServe itself - it's very near beta stage! Oh yes :) So, if you've been waiting for a stable version of TuxServe to use - your wait is almost over. On the whole, everything works - even the queues system. There are, however, still some bugs that need to be sorted out before i will release it as a beta.

If you want to check it out now, go right ahead (it's on the CVS) - but please note; there are a LOT of debug codes that are thrown out when a DCC send completes/fails - this is temporary and will be removed in proper releases (if I remember - unlike last time!).

So, stay tuned for news of the first beta release. I'm not going to give a date, incase something comes up, or if I can't defeat certain bugs (if you want to help with the bugs, checkout the CVS code - the details of current bugs are in the "todo" file). So, until then: enjoy.


Aypok...

15-03-04 :: v0.0.7alpha.

Yep. TuxServe v0.0.7alpha is now on the CVS repository (check the help page for info on how to access the TuxServe CVS). A lot of code has been changed over the weekend. I actually spent a lot of time on it over the last 3 days :) That's why there were no CVS/web updates then.

Anyhoo, the changes won't be noticable to users of the the fserve (that's why there is no tar ball of this release, or 0.0.6alpha) - but those who look at the code will see a large difference. All the nasty globals that were being used have been converted into structs - this not only makes it easier to read, but also works better (the original globals were a bad way to do what I was doing) :)

There are a few minor bugs fixed, as you'd expect - but, guess what! The infamous crashing-on-send-completion bug is back :( This is its 4th visit. But rest assured, it will be defeated. When it is, I will release another tar ball for you all.

Enjoy.


Aypok...

04-03-04 :: Quick Update.

Fixed the little bug with the advert timer - it's on the CVS now. Enjoy.


Aypok...

04-03-04 :: More Stuff.

Just another update to let you know that I'm still alive and coding TuxServe :) Development stopped for the passed few days due to a weird error with the timer that makes the advert display in the channel. That's been removed until I fix it (I'm working on it now). The CVS has been updated a few minutes back with a cool new feature: voiceme.

I'm sure most of you will have seen, or even used, the "!voiceme" command in IRC channel to get a voice. TuxServe can now be set to respond to these and voice all fserves in the channel. You can set this feature on/off, and can be given a list of upto 16 channels to work in (some people won't want it to respond in all channels). As far as I can see (and I've tested this feature quite a bit), it's all working fine. But let me know if you find a bug :)

What else? Not sure if that screw-up-the-queues thing is fixed. I think it is, but i haven't had a chance to test it yet. The updated version is on the CVS, so try for yourself :)

That'll do or now, I guess. Enjoy.


Aypok...

27-02-04 :: Yay!

We're back in business! The site and CVS have both been updated. About damn time... :)

It seems that after the update of the SF systems, my password (as well as a lot of other peoples passwords) for everything (except the main site) were reset, so we couldn't log in. I just had to change my password, and I'm back in :)

If you're having trouble getting into your SF CVS/shell/web space, then change your password on the site (you can even change it to the one it currently is - it will work). That will fix it.

On with the show: "So, what's new?" you ask. "Not much :(" I say. I've been busy trying to kill a bug in it that causes the queues not to send after a send finishes. I think it's still there at the moment, but I haven't given up on it. The bug will be fixed one way or another. *evil grin*

There are now 3 settings for the welcome message: off, random fortune quote (now displayed correctly), and read a custom welcome message from the "welcome" file (which must be in the root of the TuxServe directory, with the config, sends, and queues files).

Some of the functions now print the info to the right window, instead of just the current window. Soon all the functions will be fixed to do this.

It also seems that I have some competition in the "XChat fserve" business :) It's author is one of the dudes from the IRC network I live on. Development on it has only recently begun, so there's no code available for it yet - but I'll provide you all a link to it as soon as it's available.

That's about it so far... Back to working on that annoying queues bug. Enjoy.


Aypok...

22-02-04 :: Not My Fault...

I still can't get into my CVS repository, shell account, or webspace on SourceForge - and apparently I'm not the only one. Looks like I'll just have to wait. When you see this page, you'll know it's been fixed :)

In the meantime, I am still coding; The user queue limits are now enforced correctly, it's a bit more helpful on startup, added "make uninstall" to the Makefile, it should no longer process fserve commands in a normal non-fserve DCC chat session, and general bugs have been fixed.

That's all for now. Enjoy.


Aypok...

19-02-04 :: Earlier Than Expected

I was planning on taking another few days break, but I just couldn't keep away. Hehe. Anyhoo, so far today I managed to: correct the displaying of notes on adverts; cleaned up about half of the code that reads in the config data; and added configurable trigger restrictions.

Some parts of the functions that read in the config file were awful - even I had trouble working out what was going on - so I cleaned that up and made it a lot easier to follow (and more efficient, too). The restricted triggers bit is cool too: you can now restrict each of the triggers to those users who are voiced, hop-ed or op-ed (and above). This way, you can set triggers that only users who are voiced (and those above - eg, hops, ops, sops, owner, etc) can access - handy if you want to only distribute files to other fserves and not to the leechers.

Not bad for 1 day, I'd say :) I can't seem to update the CVS, though. I know SourceForge took it down for an hours maintenance at midday - but I still can't access it. I can't even upload this web page for you all to read, or even SSH into my SF shell. Hehe. Oh well, I'm in no rush. Enjoy.


Aypok...

18-02-04 :: Moo.

I haven't posted any new news in almost a week, so i'm just letting you know that I'm still here and TuxServe is still coming along nicely :)

Now that all the easy parts of TuxServe are coded, obviously, only the harder bits are left to do. This means that new additions will be slower to come along, and releases will be a few weeks apart. That doesn't mean I'll be doing less work on it! The CVS will still be updated as regulary as normal - don't you worry. Hehe.

But for the moment, I decided to take a few days break from coding. I plan to get back to coding on the weekend. I have still done bits here and there on this "break", but not as much as before. This also allowed me to take a step back and look at it to think "what else could I add to it?". And I came up with a few new ideas :) These will be added very soon.


Aypok...

12-02-04 :: Too Early...

Moo-ning everyone... *yawn* It's not even 10am yet, and I've released another version of TuxServe :)

The version released last night didn't exactly have any bugs, but I forgot to take out a lot of debug codes and such that get printed when certain events occur. I rereleased it due to the large number of these codes that I left in. Most of them should now be gone, but I know there are still some left in (when a send completes, you will probably see 2 codes). Just ignore anything weird that you see - such as text saying "moo #", where # is a number - these are not bugs.

Sorry for all this messing around, but as I keep pointing out; this is still an alpha :) Enjoy.


Aypok...

11-02-04 :: TuxServe v0.0.5alpha Released!

As I promised, v0.0.5alpha has been released today. It was a close call though - I nearly delayed it. I found a new bug last night that took me until midday today to fix. A very weird problem that, even now, makes no sense. But, what matters is that it's fixed :D

To get yourself a copy, browse on over to the downloads section, which is it where it lives. A few quick words about this release. I have done a lot more testing with this version (thank again to everyone who helped - they're listed in the readme file), and it does appear to work. It now sends files without crashing, too :) A few minor problems still exist (I thought it best to release this in an almost-working state, rather than postpone it even longer), such as; sometimes users can queue as many files as they like, even if the queue limit is small; sends are not removed from the sends list when completed. There are some other minor bugs/problems too. Read the readme included with the release for more information.

As per usual, if you find any bugs not listed in the "known bugs" section of the readme, then please let me know via the usual methods. I'm greatful for all the help I get with this :) Enjoy.


Aypok...

10-02-04 :: Stuff, Things, And Doohickies

No coding today - I've been out. But I've fixed the downloads page on this site. Thanks to "tass_" for letting me know. I forgot to write that page... doh. But, that's all sorted now. Yay! And such.

At the moment there is only 1 download on there - the initial TuxServe release. But that will change tomorrow morning. TuxServe v0.0.5alpha will be out. Yes indeedy. I was planning to release it on Sunday, but I got coding on it and though "after this bug fix" and "after I finish this bit of code". As you can see, I never got around to it. Until now :) This version shouldn't have such crippling problems as the last one. I hope... Hehe. Anymoo, enjoy.


Aypok...

09-02-04 :: More Fixes

Today and yesterday has seen some improvements and bugs being slaughtered in the advert department. The main problem was that adverts over about 480 characters long (including colour and formatting codes) made XChat crash. This has been fixed by a system whereby if the advert exceeds this limit, it gets split into two adverts. It does it nicely too - it doesn't split sections (that'd look messy).

In theory, this fix gives you upto almost a 1000 char advert limit! Probably a lot more than some of you will need, but quite a few people will go over the 480 character limit, so this was necessary. If needed, I can up the limit to about 1500 characters - although, I expect an advert that length would annoy a few people :) Enjoy.


Aypok...

07-02-04 :: More Bug Fixes, And Such

Now featuring in your local SourceForge CVS repository: TuxServe! I just updated it and decided to let you know what's new. Well, first up is the idle timeout system - that's been included for a long time (if you've looked at the code, you'll know), but it was always disabled due to it causing problems. All of that has now been fixed and the idle system works as well as any other fserves (it also gives a 30 second warning before it kicks the user for being idle).

I've also added in proper handling for failed and aborted DCC chat sessions, as well as proper handling for failed and aborted DCC sends (there may still be bugs, but none that I can see yet - let me know if you find any). A little bug where a completed send wasn't removed from the sends list has been fixed, too.

Recently, someone (who shall remain nameless...) commented that the "fileio.c", which was over 1000 lines long, was too long. I agree - it was. I just never noticed how big it had gotten :) To solve this, I created a new file "config.c" which handles all the file I/O that's associated with the config file. It makes no difference at all to the end user, but it should help out hackers and people who are just reading it out of boredom :)

I finally got around to creating a Makefile. See the help section for more info.

What else? Hmm... You can now actually use the non-ctcp triggers system :) It worked before, but it just ignored all non-ctcp triggers. This was sorted out in last nights CVS update, but I thought I'd mention it here.

If I seem to be coding less today and yesterday - it's because someone pointed me towards ATanks... Seriously, if you like Worms or Tanks; get this NOW! It's brilliant :) http://atanks.sourceforge.net/ (you'll need Allegro to run it) - and I /think/ it runs on most platforms - not just Unices. Even cooler. So, until next time...


Aypok...

06-02-04 :: Yay! Crashing Fixed

I fixed the `crashing when a send completes` bug about an hour ago. It's now on the CVS repository too, so you can all go get it. Saying that, it's still far from complete, and does have other bugs too :) All will be fixed in due time. Enjoy...


Aypok...

05-02-04 :: New Bug

A new bug has been found in TuxServe (in the release and the CVS): It crashes when a send completes. Which renders it kind of useless. Hehe. But hey - it's an alpha :) As you expect, I'm working on a fix now, and it'll be posted to the CVS repository as soon as it's done (as well as a post here to let you all know).

It's a bugger to find these things, since you cannot just DCC to yourself and check. I've had to rely on some very helpful friends (who I've thanked in the readme file). This has slowed down my progress somewhat, but it's not detered me :) Anymoo, back to coding for me...


Aypok...

04-02-04 :: Updates

As I promised, I'm updating the TuxServe CVS. With the latest update there are a couple of new features such as; the "on/off" feature, allowing you to switch it on and off (duh); displaying of adverts publically (as well as the original method of typing "!list") on a timer; each channel can have a different advert with different settings show (as well as having it's own triggers); plus more.

Also, I've finally included a changelog! I said that I would. It goes back to the very first versions from when I had just started coding it, right up to the last addition about 20 minutes ago :) I keep this fully updated, and is a good guide as to what new features are included in each update.

I will also be adding extra docs on hacking it and a better readme (the readme may be a week or 2 away, so don't hold your breath). Also a makefile is to be added. I'll try and have one with the next CVS update. But for now, enjoy :)


Aypok...

03-02-04 :: Problems?

Okay, sorted out all the CVS business. It was due to my rather weird file naming scheme - putting the version number at the end of the filename, eg; "main-0.0.4.c". That and me not being used to how CVS works. All file names are now simple, eg; "main.c" - the version of the file is kept via CVS. Enjoy :)


Aypok...

03-02-04 :: Problems

I know I said that some things will be fixed in the CVS version by now, but unfortuantily, there is a problem. Not with the code, that's all fine - but the CVS. Well, to be more specific - probably my lack of knowledge on how to update the CVS :) So, for the mean time, you'll have to use the crappy release code - sorry. I'm trying to work out how to upload the latest code to CVS (with the SF help pages) as we speak... I'll post again as soon as I sort it out.


Aypok...

01-02-04 :: TuxServe v0.0.4alpha Released!

That's right, the first proper release is out. Although it is still in an early stage, and I was feeling too lazy to even include a makefile - but I do include written instructions on how to compile and install it. The readme is minimal, and there are no other docs - but the online help provided in TuxServe itself should be enough to get you started. Better docs will be included with later releases and on this site. I promise. When I get time...

As you see, I took a break from coding to Gimp myself some logos and graphics. To be honest, they're a lot better than I thought they'd be. It was my first go with using Gimp 1.3 (I was used to 1.2 something)... Anyhoo...

This release includes a basic working fserve system for XChat 2. The fserve works (as far as I can see - but it may still have bugs. Let me know!) but is lacking a few minor features, such as; there is no limit on how long users can browse the fserve for; it doesn't record the stats (fixed in the CVS version). As I said - they're just minor. I decided to release early (as SourceForge says to), but I waited until I had at least a partly working system to give out :)

This site is still under construction, much like TuxServe, but this has a lower priority on my list than TuxServe does. Thusly, this site may not be updated/improved as quickly as I'd like. But I'm sure you'd rather I spent my time coding TuxServe :)

Keep checking back for new versions - I plan to release a new version every week (provided that there is a significant improvement). I'll try and update the CVS daily.

But for now, enjoy :)


Aypok...

Show only the 3 latest news articles


TuxServe and all of this site is (C)opyright 2004-2005 Aypok. Best viewed with your eyes, at 1024x768. Built with The Gimp and Screem on Slackware Linux. Best viewed in Opera. All hail the mighty Tux!