Yeah, it really is.
So please, stop doing risky stuff and fix bugs instead

mkay, time to rally some testers i guess. i'll try to find some time to commit a fix or two.

Whats left to do and would it be safe to say within the next month a release is likely to occur?

for AIs, there seem to be quite a few errors on windwos still.
i can not debug them, casue my cross compiles and on windows compiles do not work, or not compile, and the buildbot does not yet translate stack traces for the new AIs (the problem is, they all are called SkirmishAI.dll, and just differ in their path).
i have about 3 different error reports, which all seem to happen on windows only (and are AI only).

Severe bugs, or only some annoyances? Does it affect all of them, or do some work?

affects all AIs, native legacy and Java.
I can not give much details... got some reports, mainly by LoidThanead. I did some short tests only, cause it is not really fun to test if you can not get info about crashes anyway, but i would say the problems are severe enough to not yet release because of them.

I tried compiling MinGW with the script from here:
http://www.profv.de/mingw_cross_env/
but was not yet able to run it through. Ubuntu has lib PCRE (perl compatible regular expressions) without Unicode support; this script needs it with unicode support. i compiled it myself, but was not able to get the script to use the self compiled one.

an other things, that is not really release related:
it is still not possible to upload files to the wiki.


compiling with scons does not work, jk said he will do changes soon to fix that.

all ends i should work on are kind of dead, so i am not very motivated to do anything.
of course you could argue that AIs are not so important, and unsynced anyway, so we could probably still make a bug fix release later, so it is possible. but as AIs are my main concern, i would vote for not releasing yet.

There's still the incompatible util.c and util.h for visual studio, once that's fixed I can safely make a stab at debugging the above problems from VS2008, and an start a release cycle.

Either way stack traces would also be an invaluable tool that someone should really sort out

As a side note, AI developers cna use exception handling AI side to mitigate this problem and instead log crashes to their own logging systems then invoke an un-handled crash of some sort to make spring crash or simply stop processing commands. NTai did this and I think another AI did it too, but I dont remember enough to comment further on that

AF util.c and util.h still work for me, and always have.

Your using 2005 right?

2008.

*scratches head*

You might enlighten us whats broken?
And what has PCRE to do with AIs crashing?

On Windows, Java AIs have a tendency to crash without any reason at all. Also, replays with bots don't work (already on mantis.) Not sure what's more.

i cant get any development env. runnning on windows with working stack traces for AIs. so i though, maybe it will work if i crosscompile with the self compiled MinGW with the script you linked me to. this fails because of PCRE.
... and theres the problem with Java AIs aswell (which migth be related, or not) that imbaczek mentioned.
i am not very motivated these days, and spending even more days just to work around bugs that are not really my business (i feel so) to get to a testing env i could use...
sorry, i jsut wont do it, am too .. frustrated somehow.

woo hoo and I can help this time!