5.0 Release Checklist
-
- Lead Network Developer
- Posts: 2560
- Joined: Sun Jan 12, 2003 9:13 am
- Location: Palo Alto CA
- Contact:
5.0 Release Checklist
Here's what I have in my personal checklist.
Anyone got anything you want to add?
I think we can manage a release (or at least a beta?) this week.
==== SOUND ====
* Cockpit Forward/backward effects are annoying... take them out?
* Allow users to decrease sound volume.
* Disable pthread errorcheck mode.
==== PYTHON ====
* Remove extra debug output to reduce stdout size? (remove debug() function)
==== GRAPHICS ====
* Safemode's max texture size fix.
* Fix Low and Medium detail levels after the max texture size fix.
==== GAMEPLAY ====
* Reduce number of ships launched by python?
Anyone got anything you want to add?
I think we can manage a release (or at least a beta?) this week.
==== SOUND ====
* Cockpit Forward/backward effects are annoying... take them out?
* Allow users to decrease sound volume.
* Disable pthread errorcheck mode.
==== PYTHON ====
* Remove extra debug output to reduce stdout size? (remove debug() function)
==== GRAPHICS ====
* Safemode's max texture size fix.
* Fix Low and Medium detail levels after the max texture size fix.
==== GAMEPLAY ====
* Reduce number of ships launched by python?
-
- Developer
- Posts: 2150
- Joined: Mon Apr 23, 2007 1:17 am
- Location: Pennsylvania
- Contact:
are these things that need to be done or things that may already be done? Because i'm unaware of any outstanding issues with max texture size ... The texture size scaling down is done by simply choosing the next mipmap. The only problem arrises from still displaying those lower resolution files at a high resolution. The scaling back up causes significant blurriness.
This can be fixed by simply making max texture size depnedent on the chosen resolution. Low resolutions geta lower max texture size, higher resolutions get a larger one.
That seems the most logical route to take.
You can add to that checklist
==== Data ====
* Either fix or remove units that are incomplete, broken or otherwise non-functional ... can be re-enabled when issues are resolved.
This can be fixed by simply making max texture size depnedent on the chosen resolution. Low resolutions geta lower max texture size, higher resolutions get a larger one.
That seems the most logical route to take.
You can add to that checklist
==== Data ====
* Either fix or remove units that are incomplete, broken or otherwise non-functional ... can be re-enabled when issues are resolved.
-
- Lead Network Developer
- Posts: 2560
- Joined: Sun Jan 12, 2003 9:13 am
- Location: Palo Alto CA
- Contact:
Yeah I was confused because of the comment in this thread...
http://vegastrike.sourceforge.net/forum ... 8425#88425
Even if that is done, it seems there are still some problems worse graphics cards, especially with jump points it seems.
As to broken units, which ones are those? I'm going to quickly hack up a python script to launch one of each unit.
http://vegastrike.sourceforge.net/forum ... 8425#88425
Even if that is done, it seems there are still some problems worse graphics cards, especially with jump points it seems.
As to broken units, which ones are those? I'm going to quickly hack up a python script to launch one of each unit.
-
- Merchant
- Posts: 41
- Joined: Wed Jul 06, 2005 2:23 pm
- Location: In the pilot seat of my Areus.
- Contact:
I haven't been able to play the most recent SVN revisions (see first point below), so I don't know what, if anything, of this has been implemented, but here goes anyway.
==== PORTABILITY ====
* Ensure stability across platforms and configurations. (The current SVN segfaults/dies just after startup on my amd64 .)
==== SOUND ====
* Allow changing volume inside bases and with other interfaces open.
==== GAMEPLAY ====
* Ensure that ships which aren't buyable function perfectly when master_part_list.csv/the save file is hacked to obtain them. (E.g. I recall a Tesla that I'd hacked into master_part_list.csv having negative upgrade space when I bought it in some SVN revision.)
* Make more ships buyable? If the above is done, that'll be a piece of cake.
* Implement Chuck's stateless autopilot.
* Make sure everything's balanced, both in the trade system and with ship power (as well as any other areas that may need work, of course).
Again, I don't know which of these points have already been implemented, though as far as I'm concerned, the first hasn't.
Anyway, 0.5 is a big milestone, so I think it's important to round up those open ends and smooth everything out. I'd rather have it released a bit later but functioning more or less immaculately than have a rushed release that's riddled with bugs and broken features.
==== PORTABILITY ====
* Ensure stability across platforms and configurations. (The current SVN segfaults/dies just after startup on my amd64 .)
==== SOUND ====
* Allow changing volume inside bases and with other interfaces open.
==== GAMEPLAY ====
* Ensure that ships which aren't buyable function perfectly when master_part_list.csv/the save file is hacked to obtain them. (E.g. I recall a Tesla that I'd hacked into master_part_list.csv having negative upgrade space when I bought it in some SVN revision.)
* Make more ships buyable? If the above is done, that'll be a piece of cake.
* Implement Chuck's stateless autopilot.
* Make sure everything's balanced, both in the trade system and with ship power (as well as any other areas that may need work, of course).
Again, I don't know which of these points have already been implemented, though as far as I'm concerned, the first hasn't.
Anyway, 0.5 is a big milestone, so I think it's important to round up those open ends and smooth everything out. I'd rather have it released a bit later but functioning more or less immaculately than have a rushed release that's riddled with bugs and broken features.
-
- Developer
- Posts: 3980
- Joined: Fri Jan 03, 2003 4:53 am
- Location: Stanford, CA
- Contact:
link---can you go ahead and post the output and do a backtrace when it does a segfault
change vegastrike.config to not be fullscreen
then do
gdb /path/to/vegastrike
run
backtrace full
Most of the devs use amd64 linux--and so far everyone has had their troubles resolved by using the right drivers, etc...so you're right now the exception---lets get it working on your system
change vegastrike.config to not be fullscreen
then do
gdb /path/to/vegastrike
run
backtrace full
Most of the devs use amd64 linux--and so far everyone has had their troubles resolved by using the right drivers, etc...so you're right now the exception---lets get it working on your system
Vega Strike Lead Developer
http://vegastrike.sourceforge.net/
http://vegastrike.sourceforge.net/
-
- Merchant
- Posts: 41
- Joined: Wed Jul 06, 2005 2:23 pm
- Location: In the pilot seat of my Areus.
- Contact:
Heh, just recompiled it with --enable-debug and it ran, and just as I was about to post, it crashed. I managed to load a save file which I'd created in an earlier run that lasted for about as long as this one (about a minute and a half, I'd say).
Backtrace:
Full 4354 lines of output, in case it provides any more information: http://users.penguindevelopment.org/Lin ... 2.full.log
Looks like it's audio-related. I'll provide more information as necessary.
Backtrace:
Code: Select all
(gdb) backtrace full
#0 0x0000000000dd71fb in blutLoadWAVMemory ()
No symbol table info available.
#1 0x0000000000dd78ce in AUDLoadSoundFile ()
No symbol table info available.
#2 0x0000000000e9c147 in Muzak::readerThread ()
No symbol table info available.
#3 0x00002b8c6a864135 in start_thread (arg=<value optimized out>) at pthread_create.c:296
__res = <value optimized out>
pd = (struct pthread *) 0x41878940
unwind_buf = {cancel_jmp_buf = {{jmp_buf = {1099401536, -69191645264800331, 47882082599024, 47882082638496, 7, 1099403264, -69191644182309707,
-69237222828190386}, mask_was_saved = 0}}, priv = {pad = {0x0, 0x0, 0x0, 0x0}, data = {prev = 0x0, cleanup = 0x0, canceltype = 0}}}
not_first_call = 0
robust = <value optimized out>
#4 0x00002b8c6bc1d43e in clone () from /lib/libc.so.6
fstab_state = {fs_fp = 0x0, fs_buffer = 0x0, fs_mntres = {mnt_fsname = 0x0, mnt_dir = 0x0, mnt_type = 0x0, mnt_opts = 0x0, mnt_freq = 0,
mnt_passno = 0}, fs_ret = {fs_spec = 0x0, fs_file = 0x0, fs_vfstype = 0x0, fs_mntops = 0x0, fs_type = 0x0, fs_freq = 0, fs_passno = 0}}
__elf_set___libc_subfreeres_element_fstab_free__ = (const void *) 0x2b8c6bc4eae0
#5 0x0000000000000000 in ?? ()
No symbol table info available.
Looks like it's audio-related. I'll provide more information as necessary.
-
- Lead Network Developer
- Posts: 2560
- Joined: Sun Jan 12, 2003 9:13 am
- Location: Palo Alto CA
- Contact:
Hmm... first of all which SVN version is this?
Make sure you are running 11608 as that may have fixed the audio crash you are talking about.
Considering you talk about the plural "the most recent SVN revisions", I suspect you might not have updated to this latest revision.
In addition, you didn't compile with debug, so the backtrace doesn't give much information. After rerunning ./configure with --enable-debug you have to do a "make clean" to recompile all of vegastrike.
In your case, if recompiling the whole bloody thing is too slow, you should be able to only add debug info for the important files:
Remove all object files in aldrv (rm -i src/aldrv/*.o) and remove the music.o file (rm -i src/cmd/music.o) and then run "make" again, and it will figure out the missing files.
Make sure you are running 11608 as that may have fixed the audio crash you are talking about.
Considering you talk about the plural "the most recent SVN revisions", I suspect you might not have updated to this latest revision.
In addition, you didn't compile with debug, so the backtrace doesn't give much information. After rerunning ./configure with --enable-debug you have to do a "make clean" to recompile all of vegastrike.
In your case, if recompiling the whole bloody thing is too slow, you should be able to only add debug info for the important files:
Remove all object files in aldrv (rm -i src/aldrv/*.o) and remove the music.o file (rm -i src/cmd/music.o) and then run "make" again, and it will figure out the missing files.
-
- Developer
- Posts: 3980
- Joined: Fri Jan 03, 2003 4:53 am
- Location: Stanford, CA
- Contact:
additionally---if the latest Svn version still breaks try without audio (in the setup util)
another thing: don't load an old savegame--it takes minutes-hours to convert them.... try making a new game by clicking campaign
we'll make that obvious in the release by only showing new savegameso
another thing: don't load an old savegame--it takes minutes-hours to convert them.... try making a new game by clicking campaign
we'll make that obvious in the release by only showing new savegameso
Vega Strike Lead Developer
http://vegastrike.sourceforge.net/
http://vegastrike.sourceforge.net/
-
- Merchant
- Posts: 41
- Joined: Wed Jul 06, 2005 2:23 pm
- Location: In the pilot seat of my Areus.
- Contact:
I was using 11608. And I did think it was weird that it was still saying there were no symbols. I'll make clean and recompile. (And considering I use Gentoo, that doesn't bother me much. I'll pick compiling Vega Strike over compiling KDE or GCC any day .)
And the old savegame was made maybe a revision earlier in the few minutes before VS crashed, so I don't think that matters much...? Restarting a savegame after every svn up -- that would be ridiculous.
Edit: okay, more results. It appears to crash while loading a custom song I've added to the playlist (it's in MP3 format, and I know for a fact that it worked in revision 10923, which is the revision I played before I tried to update a few days ago). Backtrace here; full output here.
And the old savegame was made maybe a revision earlier in the few minutes before VS crashed, so I don't think that matters much...? Restarting a savegame after every svn up -- that would be ridiculous.
Edit: okay, more results. It appears to crash while loading a custom song I've added to the playlist (it's in MP3 format, and I know for a fact that it worked in revision 10923, which is the revision I played before I tried to update a few days ago). Backtrace here; full output here.
-
- Lead Network Developer
- Posts: 2560
- Joined: Sun Jan 12, 2003 9:13 am
- Location: Palo Alto CA
- Contact:
well then that makes sense... Vega Strike does not have MP3 support because it's a lot of work for little gain (OGG is better quality).
I remember at one point we used OpenAL's built-in support for MP3's on Linux, which was somewhat buggy... maybe that's the issue.
As to the crash, perhaps it should fail a little bit more gracefully in that case.
I remember at one point we used OpenAL's built-in support for MP3's on Linux, which was somewhat buggy... maybe that's the issue.
As to the crash, perhaps it should fail a little bit more gracefully in that case.
-
- Merchant
- Posts: 41
- Joined: Wed Jul 06, 2005 2:23 pm
- Location: In the pilot seat of my Areus.
- Contact:
That explains a lot. I'll get working on converting those MP3s then (which is easy enough). But yeah, just an stderr message and an attempted song skip would be good. Of course, that could create infinite loops of skipping songs. Maybe auto-discard all entries that cannot be played, and quit attempting when there are no more valid entries.
Anyway, I'll post a new topic if something else comes up; I don't want to keep this thread hijacked any longer .
Anyway, I'll post a new topic if something else comes up; I don't want to keep this thread hijacked any longer .