[http://sourceforge.net/p/ufoai/bugs/2453 Item 2453] imported from sourceforge.net tracker on 2013-01-28 19:35:59

Ufo crashes whenever I click in the geoscape to build an additional installation (after the HQ, e.g. an Ufo yard or a SAM base), saying 'Aborted.' in the console.
Please give me a hint about how to provide additional info of this, if there's a log or someting I didn't manage to find yet.
I was using r29497.
HTH,
Gerwulf
===== Comments Ported from Sourceforge =====

====== gerwulf (2010-04-21 07:50:07) ======

FYI: I did a 'make clean', a new 'configure' run and removed my entire ~/.ufoai folder to make sure no side effects would cause this, but the behaviour stays the same. Additionally: Ubuntu 10.04 (up to date) on a dual core intel CPU, with an ATI HD3400 card running on the fglrx driver.
====== gerwulf (2010-04-21 07:54:41) ======


====== gerwulf (2010-04-21 07:55:19) ======


====== geever (2010-04-21 10:33:41) ======

I cannot reproduce it, please paste the "abort" message (and the last few lines of the console output here, your ufoconsole. doesn't contain any info about the crash. :(

-geever
====== gerwulf (2010-04-21 15:14:58) ======

Here you go (but besides of that 'aborted' message, it's just the same (still happens in r29502):
>>>
...nations: 8
...cities: 457

Sanity check for script data
...buildings ok
...tech ok
...aircraft ok
CP_ItemsSanityCheck: Item craft_ufo_scout has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_fighter has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_harvester has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_bomber has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_ripper has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_mothership has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_gunboat has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_corrupter has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_carrier has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_supply has zero size set.
...items failed
...items ok
NAT_ScriptSanityCheck Errors: 0
...nations ok
music change to karlmacklin_geoscape (from van_theme)
Aborted
bsnippe@bsni-glx:~/ufoai/trunk$
<<<

Is there something like a &#039;verbose&#039; switch for more error logging?
====== tlh2000 (2010-04-21 15:58:57) ======

aborted is a segfault - please post a backtrace from gdb (see wiki debugging article) http://ufoai.ninex.info/wiki/index.php/Debugging#gdb
====== gerwulf (2010-04-21 16:29:11) ======

Okay, looks like the error is connected to the direct draw part of my fglrx driver. I remember that I had troubles getting direct draw to work under Ubuntu 10.04 Beta2 after I upgraded from 9.10, but a fresh install of 10.04 Beta2 fixed this.
Maybe an upcoming version of fglrx fixes this bug, but then again, why did it work until yesterday? I do not remember an update to fglrx in the past few days, but I&#039;ll check anyway.

>>>
bsnippe@bsni-glx:~/ufoai/trunk$ gdb ufo
GNU gdb (GDB) 7.1-ubuntu
Copyright (C) 2010 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "i486-linux-gnu".
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>...
Reading symbols from /home/bsnippe/ufoai/trunk/ufo...done.
(gdb) run +set vid_fullscreen 0 +set vid_grabmouse 0 +set developer 1
Starting program: /home/bsnippe/ufoai/trunk/ufo +set vid_fullscreen 0 +set vid_grabmouse 0 +set developer 1
[Thread debugging using libthread_db enabled]

---- filesystem initialization -----
Adding game dir: ./base
Adding game dir: /home/bsnippe/.ufoai/2.3-dev/base

[...]

CP_ItemsSanityCheck: Item craft_ufo_carrier has zero size set.
CP_ItemsSanityCheck: Item craft_ufo_supply has zero size set.
...items failed
...items ok
NAT_ScriptSanityCheck Errors: 0
...nations ok
music change to karlmacklin_geoscape (from van_theme)

Program received signal SIGSEGV, Segmentation fault.
0x02586edd in ?? () from /usr/lib/fglrx/dri/fglrx_dri.so
(gdb) bt full
#0 0x02586edd in ?? () from /usr/lib/fglrx/dri/fglrx_dri.so
No symbol table info available.
#1 0x00000000 in ?? ()
No symbol table info available.
(gdb)
<<<


====== gerwulf (2010-04-21 16:30:36) ======

Btw, to make extra sure, I removed my entire working copy, did a fresh checkout from svn, but the error stays. :-(
====== gerwulf (2010-04-21 16:34:34) ======

FYI: Launchpad tells me that the last update to the fglrx package in Ubuntu 10.04 happened: 2010-04-15 16:14:25 UTC.
====== gerwulf (2010-04-22 14:23:22) ======

Okay, I did some testing, went back in time (step by step, up to r29475), but the results seemed incoherent, so I updated to the head revision again.
But now, it looks I have found a singularity (and a way for me to play at all ;-) ):
If I start Ufo, begin a new campaign, build the HQ and then an external installation, it crashes (with the segfault in fglrx_dri.so mentioned earlier).
If I start Ufo, begin a new campaign, build the HQ, save my game, then create an external installation, it still crashes.
--> But: if I start Ufo, load that saved game (from the last try, so my HQ is already built), and then create an external installation, it just works as it should.
Summary: the crash happens only if I start a fresh campaign, not if I load a saved game. And yes, I did delete my user settings before testing and did not load a save game from an older/newer revision.
====== tlh2000 (2010-04-22 14:40:16) ======

maybe related to the Xorg bug that was introduced in 10.04?

https://lists.ubuntu.com/archives/ubuntu-devel/2010-April/030673.html
====== gerwulf (2010-04-22 15:38:58) ======

According to the wiki pages related to that bug, ATI cards are not affected: "This does not affect cards using proprietary drivers or not using DRI2 because it is specific to the glx module that the open drivers use. Intel will always be affected since DRI2 is used with and without KMS, ATI uses DRI1 without KMS."
The crash seems to happen when the geoscape screen wants to show the installation dialog (where you can set it&#039;s name). It escapes me how this call could be different depending on if the game was loaded from a save game or not, but I don&#039;t know exactly how to test this. Maybe I could set a break point there an check the code, but on the other hand, I&#039;m not too familiar with C/C++ debugging under Linux. :-/
====== geever (2010-05-01 20:49:21) ======

Try disabling postprocessing and maybe GLSL shaders.

I don&#039;t know if we can help you more. It seems not to be a problem with UFO:AI.

-geever
====== gerwulf (2010-05-06 15:16:42) ======

I even tried disabling all graphic options, yet no success. Nevermind, since I have that workaround via the save game, it&#039;s not a show stopper for me. And I agree, it&#039;s probably not a fault with UFO:AI, maybe rather an unlucky graphics driver quirk. Setting this as invallid, since it can&#039;t be fixed in UFO:AI.