[http://sourceforge.net/p/ufoai/bugs/385 Item 385] imported from sourceforge.net tracker on 2013-01-28 18:18:28

Sending me back to desktop with this massage :

SZ_GetSpace: overflow without allowoverflow set

version 2.0 RC5
I have saved to all boxes and trying to save over them.
One time of the five saves.
===== Comments Ported from Sourceforge =====

====== deanhouse (2006-09-30 06:47:43) ======

Logged In: YES
user_id=752750

I am having the same problem on Windows XP. I can no
longer progress in my game as I cannot save it to any box
without it crashing!
====== cassiterite (2006-09-30 08:31:24) ======

Logged In: YES
user_id=1171370

Is this only happening after all saved-game slots are filled
or does something else trigger it (e.g. a certain point in
the game plot-line?)

This would appear be a major bug so I've upgraded it's
priority. Please provide as many specific details as
possible so we can try and reproduce it and find a solution.
====== nobody (2006-09-30 10:12:53) ======

Logged In: NO

I have all saved-game slots are filled (1-8) and then
started the problem. Samehow i have succeed to save in slot
1 again but only ones. I have won 6 missions and olny in the
last mission i have a actor killed. I replase this actor and
produced one armor (kevlar) and equip this man. Then the
final moment before next mission (always save before) start
the problem. Researching "Advanced Plasma Control"

====== deanhouse (2006-09-30 15:57:33) ======

Logged In: YES
user_id=752750

I have NOT filled all the game slots. It seems like I can
progress up to a point in the game (never even into
February) before my game becomes unsavable. Also when I
try to save to an existing slot, it corrupts that new slot
so if I try to load it, then IT will immediately crash.

I had previously won 6-8 missions and haven't lost any
soldiers. It seems I can continue playing as far as I
want but if I ever go to save it, the game crashes and it
doesn't complete the save, so I am forced to play from the
last good save point.

I can even load the last working save point then
IMMEDIATELY go and save and it'll crash.
====== enzo (2006-09-30 16:59:48) ======

Logged In: YES
user_id=87262

The same just happened to me while trying to save after the
Wellington mission.

I had only one slot filled before and tried to overwrite it.

The problem doesn't seem to occur in 2.1-dev1 by the way.

====== cassiterite (2006-09-30 17:09:37) ======

Logged In: YES
user_id=1171370

Can anyone with this problem please upload their
"ufoconsole.log" after getting this error (I think this log
is located in the "base" directory... eg on Linux it is
under ~/.ufoai/base, but somewhere else on Windows).

Also try running the game from the command line as "ufo +set
vid_fullscreen 0 +set vid_grabmouse 0 +set developer 1" to
try and capture more debugging output which could be useful.
====== thomas_k_70 (2006-09-30 17:28:34) ======

Logged In: YES
user_id=1610203

Sorry for the length but don't now how to submit a file

I have started the bug report and i am now at this ...
please don't shoot the messenger

Winsock Initialized
Console initialized.
------- Loading ref_gl.dll -------
ref_gl version: GL 0.12
Initializing OpenGL display
...setting mode 4: 800 600 W
...setting windowed mode
GL PFD: color(32-bits) Z(24-bit)
Desktop resolution: 1024:768
GL_VENDOR: ATI Technologies Inc.
GL_RENDERER: RADEON 9250/9200 Series DDR x86/MMX/3DNow!/SSE
GL_VERSION: 1.3.1072 WinXP Release
GL_EXTENSIONS: GL_ARB_multitexture GL_EXT_texture_env_add
GL_EXT_compiled_vertex_array GL_S3_s3tc
GL_ARB_occlusion_query GL_ARB_point_parameters
GL_ARB_texture_border_clamp GL_ARB_texture_compression
GL_ARB_texture_cube_map GL_ARB_texture_env_add
GL_ARB_texture_env_combine GL_ARB_texture_env_crossbar
GL_ARB_texture_env_dot3 GL_ARB_texture_mirrored_repeat
GL_ARB_transpose_matrix GL_ARB_vertex_blend
GL_ARB_vertex_buffer_object GL_ARB_vertex_program
GL_ARB_window_pos GL_ATI_element_array GL_ATI_envmap_bumpmap
GL_ATI_fragment_shader GL_ATI_map_object_buffer
GL_ATI_texture_env_combine3 GL_ATI_texture_mirror_once
GL_ATI_vertex_array_object GL_ATI_vertex_attrib_array_object
GL_ATI_vertex_streams GL_ATIX_texture_env_combine3
GL_ATIX_texture_env_route
GL_ATIX_vertex_shader_output_point_size GL_EXT_abgr
GL_EXT_bgra GL_EXT_blend_color GL_EXT_blend_func_separate
GL_EXT_blend_minmax GL_EXT_blend_subtract
GL_EXT_clip_volume_hint GL_EXT_draw_range_elements
GL_EXT_fog_coord GL_EXT_multi_draw_arrays
GL_EXT_packed_pixels GL_EXT_point_parameters
GL_EXT_rescale_normal GL_EXT_secondary_color
GL_EXT_separate_specular_color GL_EXT_stencil_wrap
GL_EXT_texgen_reflection GL_EXT_texture3D
GL_EXT_texture_compression_s3tc GL_EXT_texture_cube_map
GL_EXT_texture_edge_clamp GL_EXT_texture_env_combine
GL_EXT_texture_env_dot3 GL_EXT_texture_filter_anisotropic
GL_EXT_texture_lod_bias GL_EXT_texture_mirror_clamp
GL_EXT_texture_object GL_EXT_texture_rectangle
GL_EXT_vertex_array GL_EXT_vertex_shader
GL_HP_occlusion_test GL_NV_blend_square
GL_NV_occlusion_query GL_NV_texgen_reflection
GL_SGI_color_matrix GL_SGIS_generate_mipmap
GL_SGIS_multitexture GL_SGIS_texture_border_clamp
GL_SGIS_texture_edge_clamp GL_SGIS_texture_lod
GL_SUN_multi_draw_arrays GL_WIN_swap_hint
WGL_EXT_extensions_string WGL_EXT_swap_control
...enabling GL_EXT_compiled_vertex_array
...enabling WGL_EXT_swap_control
...using GL_EXT_point_parameters
...using GL_ARB_multitexture
...using GL_EXT_texture_env_combine
...GL_SGIS_multitexture deprecated in favor of ARB_multitexture
...using GL_ARB_texture_compression
...using GL_EXT_texture_filter_anisotropic [16 max] [ 1
selected]
...using GL_EXT_texture_lod_bias
...using GL_EXT_stencil_wrap
...using GL_EXT_fog_coord
...GL_ARB_fragment_program not found
...GL_ATI_separate_stencil not found
...max texture size:
......detected 2048
...SDL_ttf version 2.0.7 - we need at least 2.0.7
...SDL_ttf inited
------------------------------------
------- Loading ref_gl.dll -------
ref_gl version: GL 0.12
Initializing OpenGL display
...setting mode 4: 800 600 W
...setting windowed mode
GL PFD: color(32-bits) Z(24-bit)
Desktop resolution: 1024:768
GL_VENDOR: ATI Technologies Inc.
GL_RENDERER: RADEON 9250/9200 Series DDR x86/MMX/3DNow!/SSE
GL_VERSION: 1.3.1072 WinXP Release
GL_EXTENSIONS: GL_ARB_multitexture GL_EXT_texture_env_add
GL_EXT_compiled_vertex_array GL_S3_s3tc
GL_ARB_occlusion_query GL_ARB_point_parameters
GL_ARB_texture_border_clamp GL_ARB_texture_compression
GL_ARB_texture_cube_map GL_ARB_texture_env_add
GL_ARB_texture_env_combine GL_ARB_texture_env_crossbar
GL_ARB_texture_env_dot3 GL_ARB_texture_mirrored_repeat
GL_ARB_transpose_matrix GL_ARB_vertex_blend
GL_ARB_vertex_buffer_object GL_ARB_vertex_program
GL_ARB_window_pos GL_ATI_element_array GL_ATI_envmap_bumpmap
GL_ATI_fragment_shader GL_ATI_map_object_buffer
GL_ATI_texture_env_combine3 GL_ATI_texture_mirror_once
GL_ATI_vertex_array_object GL_ATI_vertex_attrib_array_object
GL_ATI_vertex_streams GL_ATIX_texture_env_combine3
GL_ATIX_texture_env_route
GL_ATIX_vertex_shader_output_point_size GL_EXT_abgr
GL_EXT_bgra GL_EXT_blend_color GL_EXT_blend_func_separate
GL_EXT_blend_minmax GL_EXT_blend_subtract
GL_EXT_clip_volume_hint GL_EXT_draw_range_elements
GL_EXT_fog_coord GL_EXT_multi_draw_arrays
GL_EXT_packed_pixels GL_EXT_point_parameters
GL_EXT_rescale_normal GL_EXT_secondary_color
GL_EXT_separate_specular_color GL_EXT_stencil_wrap
GL_EXT_texgen_reflection GL_EXT_texture3D
GL_EXT_texture_compression_s3tc GL_EXT_texture_cube_map
GL_EXT_texture_edge_clamp GL_EXT_texture_env_combine
GL_EXT_texture_env_dot3 GL_EXT_texture_filter_anisotropic
GL_EXT_texture_lod_bias GL_EXT_texture_mirror_clamp
GL_EXT_texture_object GL_EXT_texture_rectangle
GL_EXT_vertex_array GL_EXT_vertex_shader
GL_HP_occlusion_test GL_NV_blend_square
GL_NV_occlusion_query GL_NV_texgen_reflection
GL_SGI_color_matrix GL_SGIS_generate_mipmap
GL_SGIS_multitexture GL_SGIS_texture_border_clamp
GL_SGIS_texture_edge_clamp GL_SGIS_texture_lod
GL_SUN_multi_draw_arrays GL_WIN_swap_hint
WGL_EXT_extensions_string WGL_EXT_swap_control
...enabling GL_EXT_compiled_vertex_array
...enabling WGL_EXT_swap_control
...using GL_EXT_point_parameters
...using GL_ARB_multitexture
...using GL_EXT_texture_env_combine
...GL_SGIS_multitexture deprecated in favor of ARB_multitexture
...using GL_ARB_texture_compression
...using GL_EXT_texture_filter_anisotropic [16 max] [ 1
selected]
...using GL_EXT_texture_lod_bias
...using GL_EXT_stencil_wrap
...using GL_EXT_fog_coord
...GL_ARB_fragment_program not found
...GL_ATI_separate_stencil not found
...max texture size:
......detected 2048
...SDL_ttf version 2.0.7 - we need at least 2.0.7
...SDL_ttf inited
...registering 0 fonts
------------------------------------

------- sound initialization -------
Loading snd_sdl sound driver
Soundsystem: SDL.
SDL version: 1.2.8
SDL audio driver is "dsound".
Bits: 16
Frequency: 48000
Samples: 4096
Channels: 2
sound sampling rate: 48000
CD Audio Initialized
execing autoexec.cfg
"version" is "UFO: Alien Invasion 2.0-RC5 x86 Sep 26 2006
Win32 DEBUG"
...could not set to language:
...could not set to system language
Shared Client/Server Info loaded
...precaching 49 menus
Can't find pics/
====== UFO Initialized ======

Changing to Singleplayer
Savefile version 5 detected
CL_ParseNations: unknown token "pos" ignored (nation latin)
CL_ParseNations: unknown token "pos" ignored (nation states)
CL_ParseNations: unknown token "pos" ignored (nation europa)
Global data loaded - size 5593844 bytes
Activate stage intro
Campaign 'slot0' loaded.
...aircraft inited
Team './base/save/team.mpt' saved.
SDL audio device shut down.

====== deanhouse (2006-10-02 06:54:26) ======

Logged In: YES
user_id=752750

I also couldn't figure out how to upload.

[http://sourceforge.net/docman/display_doc.php?
docid=24202&group_id=1#file_attachments] I don't see this
checkbox/browse button on my bug screen.

Anyway, I tried it with the and without the extended debug
info. I saved the two log files into a zip available here:
http://www.daylightcreations.com/ufo/ufoconsole.zip

Thanks,
Dean
====== nobody (2006-10-03 12:42:57) ======

Logged In: NO

The problem seems not to be dependent of the save slots. I played, removed
the save dir. Now, after having loaded an old save it still didn´t work...
-> crash to desktop.
====== cassiterite (2006-10-03 19:33:23) ======

Logged In: YES
user_id=1171370

Ok, I haven't had time to look into this bug in detail since
it doesn't affect trunk SVN (I intend to switch to 2.0
branch to investigate soon).

It appears to me to be something related with saving
research dependencies.

First report of this bug that I know of:
http://sourceforge.net/tracker/index.php?func=detail&aid=1555677&group_id=157793&atid=805242

The log indicates that messages were stored then a crash
occurred. The campaign data is saved shortly after messages,
and included in the campaign data is research related data.

Another reference:
http://ufo.myexp.de/phpBB2/viewtopic.php?p=3679&highlight=#3679

again this forum post appears to indicate it was related to
saving research data...

I have looked through patches applied to trunk SVN which
weren't merged to 2.0 branch but have so far not identified
the change that fixed this in trunk (some major changes were
made to the research system in trunk... it might be related
to this).
====== nobody (2006-10-03 19:51:16) ======

Logged In: NO

I am experiencing crashes with the exact same crash msg.
Everything works ok up until the end of feb 2084 but saves
cause a crash when I am into march.
====== cassiterite (2006-10-03 20:49:52) ======

Logged In: YES
user_id=1171370

Ok I spent a bit more time tracking this down. It looks like
it may not be fixed in trunk after all (well theoretically
anyway).

Here were my real-time thoughts posted ot the IRC channel:

<ubequitz> actually: "Error: SZ_GetSpace: overflow without
allowoverflow set" errors will be encountered when
SZ_GetSpace fails because the space required exceeds
"buf->maxsize" ... in the case of saved games buf->maxsize
is set to MAX_GAMESAVESIZE (via a call to SZ_Init in the
gamesave code)
<ubequitz> MAX_GAMESAVESIZE is inturn defined as:
MAX_TEAMDATASIZE + 16384 + sizeof(globalData_t)
<ubequitz> where MAXTEAMDATASIZE is set to 32768 and
sizeof(globalData_t) seems to be independent of the number
of messages (and independent of "MAX_MSGLEN" in net_chan.h)
<ubequitz> so the number of messages is limited to say 16K
bytes.... which is about 200 messages (say around 75
characters per message)
<ubequitz> the question though is really... how come saved
games can be like 3MB>?
<ubequitz> is sizeof(globalData_t) reall that big?
<mattn2> yes
<mattn2> most of the savegames are zeros though
<nairan> wow i always wanted a "zero" savegame =P
<mattn2> due to the many [MAX_VAR] string arrays
<ubequitz> mattn2: I propose in the saved game code adding
the size of the message stack to the max saved game size


(apologies for the formatting)

So it looks like there is a relatively simple fix (although
the simple fix will make saved games even slightly bigger!).
Another improvement might be to reduce the spam from
messages like "drop shop low on fuel").
====== cassiterite (2006-10-04 13:00:56) ======

Logged In: YES
user_id=1171370

Thanks to everyone who helped out by posting the gory
details leading to this bug.

It should now be "fixed" in trunk SVN revision 4000. The
commit log said: client/client.h increase MAX_GAMESAVESIZE
to allow for 3MB of dynamic data (will fix sourceforge bug
1568183 for all real-world cases, although a longer-term
plan is to code a better save-game system & format).

I&#039;m closing this for now (a maximum of 3MB of messages
should be quite sufficient), although as the changelog says
a more complete fix is planned (it should be possible to
make the saved-game format more flexible, smaller and free
from a "MAX_GAMESAVESIZE" restriction).
====== enzo (2006-10-04 13:18:45) ======

Logged In: YES
user_id=87262

Thank&#039;s for fixing this, but I think this must also be fixed
in the 2.0 branch. For me, the bug never occurred in
v2.1-dev1, only in the 2.0 branch.

I also checked the save game size. A sample save game is
5465kB here mostly consisting of 0x00 bytes. I can compress
the file with bzip2 to 9kB! You should really think about a
different save game format!