Crasher a few seconds after first-rez...

Well, I'm impressed that we have a Linux client and that it runs at all, so kudos for that. :)
Sadly I'm getting a crash a few seconds after logging in and 'first rez'.

I suspect it's related to image decoding. Why? :-
1) The first few seconds after log-in, things are looking good. My av rezzes (my av is not conspicuously textured normally so no great clues there), I see prims start to rez, but in these first few seconds no texture has begun to rez, just the default SL 'square' unrezzed-texture-texture. Then - crash. My guess is just that this is the point at which the first texture gets sucked down and started to decode.
2) On subsequent logins I get an immediate crash before anything conspicuously rezzes. Clearing my cache gives me the several seconds of textureless-uncrashedness I previously enjoyed. So the guess is that in the cached case, the texture gets pulled from cache immediately, decode starts, and thus the crash is instant.

Nothing conspicuous in the log I think (it detects my 128MB video RAM as 16MB, but that's a problem noted elsewhere), but here's some highlights anyway:

2006-02-02T00:55:24Z INFO: Second Life version 1.8.2
2006-02-02T00:55:24Z INFO: Local time: 2006-02-02T00:55:24 GMT
2006-02-02T00:55:24Z INFO: CPU info:
2006-02-02T00:55:24Z processor : 0 vendor_id : AuthenticAMD cpu family : 6 model : 2 model name : AMD Athlon(tm) Processor stepping : 1 cpu MHz : 757.485 cache size : 512 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 sep mtrr pge mca cmov pat pse36 mmx fxsr syscall mmxext 3dnowext 3dnow bogomips : 1510.60
2006-02-02T00:55:24Z INFO: Memory info:
2006-02-02T00:55:24Z total: used: free: shared: buffers: cached: Mem: 396066816 292040704 104026112 0 6004736 165797888 Swap: 540286976 243945472 296341504 MemTotal: 386784 kB MemFree: 101588 kB MemShared: 0 kB Buffers: 5864 kB Cached: 131756 kB SwapCached: 30156 kB Active: 101644 kB Inactive: 88836 kB HighTotal: 0 kB HighFree: 0 kB LowTotal: 386784 kB LowFree: 101588 kB SwapTotal: 527624 kB SwapFree: 289396 kB
2006-02-02T00:55:24Z INFO: OS info: Linux 2.4.31 #16 Tue Jun 14 19:24:07 BST 2005 i686

[...]

2006-02-02T00:55:24Z WARNING: Frame buffer has less than 8 bits of alpha. Avatar texture compositing will fail.
2006-02-02T00:55:24Z INFO: Couldn't initialize mipmap generation
2006-02-02T00:55:24Z INFO: Couldn't initialize GL_EXT_paletted_texture
2006-02-02T00:55:24Z INFO: Couldn't initialize GL_NV_vertex_array_range
2006-02-02T00:55:24Z INFO: Couldn't initialize GL_NV_fence
2006-02-02T00:55:24Z INFO: Couldn't initialize separate specular color
2006-02-02T00:55:24Z INFO: Couldn't initialize anisotropic filtering
2006-02-02T00:55:24Z WARNING: Unknown feature mask NVIDIA
2006-02-02T00:55:24Z INFO: Graphics Card memory set to 16 MB
2006-02-02T00:55:24Z WARNING: SystemRoot not found, attempting to load fonts from default path.
2006-02-02T00:55:24Z WARNING: SystemRoot not found, attempting to load fonts from default path.
2006-02-02T00:55:24Z WARNING: Couldn't load font ヒラã‚*゙ノ角ゴ Pro W3.otf

[...]

2006-02-02T00:55:24Z INFO: GL_VENDOR NVIDIA Corporation
2006-02-02T00:55:24Z GL_RENDERER GeForce FX 5200/AGP/3DNOW!
2006-02-02T00:55:24Z GL_VERSION 2.0.0 NVIDIA 76.67

[...]

2006-02-02T01:11:15Z INFO: Receiving f9a270ef-3566-13ba-c831-e1f7b9aba4f4, size 77 bytes
2006-02-02T01:11:15Z INFO: Receiving 3822c656-732a-baa3-132a-7b3ee2bf9cfc, size 62 bytes
2006-02-02T01:11:15Z INFO: Receiving e7743d65-d2ab-fcff-d3cd-a9745f29b2b1, size 103 bytes
2006-02-02T01:11:15Z INFO: Receiving ce921c38-b502-2d59-ed12-2cdd944054a0, size 71 bytes
2006-02-02T01:11:16Z INFO: Rename failure (18) - /tmp/39953525-d9ba-cdb9-d3f5-066a45af7064.tmp to /dev/shm/SecondLife-linux-02012006/SecondLife/cache/mute_87905b8f-59df-44b6-8df8-1c8c88aea307.tmp
2006-02-02T01:11:16Z INFO: Copied the file - this should *NEVER* happen in production
2006-02-02T01:11:16Z INFO: xfer from 66.150.244.151:12036 complete: /dev/shm/SecondLife-linux-02012006/SecondLife/cache/mute_87905b8f-59df-44b6-8df8-1c8c88aea307.tmp
2006-02-02T01:11:16Z INFO: LLMuteList::processMuteListFile()
2006-02-02T01:11:16Z INFO: LLCircuit::addCircuitData for 72.5.12.84:13004
2006-02-02T01:11:16Z INFO: Adding new region (983:999)
2006-02-02T01:11:16Z INFO: Host: 72.5.12.84:13004
2006-02-02T01:11:16Z INFO: simulator_enable() Enabling 72.5.12.84:13004 with code 25706
2006-02-02T01:11:17Z INFO: Decoded 7 msgs, net 0 msgs queued this frame!
2006-02-02T01:11:17Z WARNING: LLWorld::addRegion for a simulator that already exists!
2006-02-02T01:11:17Z INFO: simulator_enable() Enabling 72.5.12.84:13004 with code 25706
2006-02-02T01:11:18Z INFO: Starting transfer for ea4f91b3-e53d-f20a-0fa0-eafafd94afa8
2006-02-02T01:11:18Z INFO: Starting transfer for 064981da-ada3-c7c8-715a-e04290906764
2006-02-02T01:11:18Z INFO: Starting transfer for ea7685ad-bf1c-6ded-5951-dec24199b0ba
2006-02-02T01:11:18Z INFO: Starting transfer for 9f3572c7-3a1b-8ed7-06ba-2a66e59ee951
2006-02-02T01:11:18Z WARNING: LLWorld::addRegion for a simulator that already exists!
2006-02-02T01:11:18Z INFO: simulator_enable() Enabling 72.5.12.84:13004 with code 25706
2006-02-02T01:11:18Z INFO: Decoded 4 msgs, net 0 msgs queued this frame!
2006-02-02T01:11:18Z INFO: Receiving e32e0a57-5892-77e8-052d-92a3c0563229, size 1071 bytes
2006-02-02T01:11:18Z WARNING: Out of order packet in transfer e32e0a57-5892-77e8-052d-92a3c0563229, got 1 expecting 0
2006-02-02T01:11:18Z INFO: Playing back delayed packet 1
2006-02-02T01:11:18Z INFO: Receiving 5ea12974-323a-aa35-acbe-bd3c13056839, size 748 bytes
2006-02-02T01:11:18Z INFO: Receiving 2d7b9e25-f699-7592-a345-fd099f191b6f, size 774 bytes
2006-02-02T01:11:19Z INFO: Receiving ed2ca018-021c-12e0-01ea-e96b3019e57e, size 506 bytes
2006-02-02T01:11:19Z INFO: Starting transfer for bbe05f72-8a97-b935-e914-20236ca2d685
2006-02-02T01:11:19Z INFO: Sending cached texture data
2006-02-02T01:11:19Z INFO: Receiving 16216de5-81a8-2d15-d662-e9c342d311dc, size 630 bytes
2006-02-02T01:11:20Z INFO: Kills on unknown objects: 5
2006-02-02T01:11:20Z INFO: Unknown object updates: 1
2006-02-02T01:11:20Z INFO: remove_marker_file()
Abort

Sadly I'm very doubtful that I can get a gdb backtrace, though I may be able to get an ltrace if that's helpful.

Why's that? You should be able to get one with:

LD_LIBRARY_PATH=lib gdb bin/secondlife-bin

(But whether it has helpful symbols in is another matter..)

I don't have a version of gdb compiled for glibc2.3 (2.3 isn't my 'native' glibc).

I got a backtrace - but it's full of '??', so not helpful.

Amusingly, under Valgrind I don't get a crash (so no more useful backtrace ), but of course things are glacial and it has trouble keeping connected to a sim because of enormous ping times. (Valgrind notices lots of errors - mismatched free/delete/delete[], overlapping src/dest on a memcpy, etc, but I can't say what might be relevant.)

whew - got the crash (abort) under valgrind after all. I like valgrind backtraces so much more than GDB's. :) Let me know if there's more I can dig into here.

[snip]
==5902== by 0x34958977: abort (in /snout/libc6-232/libc-2.3.2.so)
==5902== by 0x348D6576: (within /usr/local/lib/libstdc++.so.5)
==5902== by 0x348D65B3: std::terminate() (in /usr/local/lib/libstdc++.so.5)
==5902== by 0x348D6725: __cxa_throw (in /usr/local/lib/libstdc++.so.5)
==5902== by 0x346BFC53: kd_header_in::get_bits(int) (compressed_local.h:297)
==5902== by 0x346BCC53: kd_block::parse_packet_header(kd_header_in&, kd_buf_server*, int) (blocks.cpp:352)
==5902== by 0x346DA847: kd_precinct::read_packet() (compressed.cpp:2594)
==5902== by 0x346DEB97: kd_precinct::desequence_packet() (compressed_local.h:2573)
==5902== by 0x346D93C6: kdu_subband::open_block(kdu_coords, int*) (compressed_local.h:2885)
==5902== by 0x346DF7FC: kd_decoder::decode_row_of_blocks() (decoder.cpp:229)
==5902== by 0x346DF67A: kd_decoder::pull(kdu_line_buf&, bool) (decoder.cpp:185)
==5902== by 0x347042F9: kd_synthesis::horizontal_synthesis(kd_line_cosets&) (kdu_sample_processing.h:762)
==5902== by 0x34703492: kd_synthesis::pull(kdu_line_buf&, bool) (synthesis.cpp:367)
==5902== by 0x81646D4: LLJ2CDecodeState::processTileDecode(float, int) (in /var/opt/secondlife/SecondLife-linux-02012006/bin/secondlife-bin)
==5902== by 0x816265D: LLImageJ2C::decode(LLImageRaw&, float, LLImageJ2C::ECodeStreamMode, int, int) (in /var/opt/secondlife/SecondLife-linux-02012006/bin/secondlife-bin)
==5902== by 0x8162292: LLImageJ2C::decode(LLImageRaw&, float, int, int) (in /var/opt/secondlife/SecondLife-linux-02012006/bin/secondlife-bin)
==5902== by 0x98030C5: LLViewerImage::decodeImage(float) (in /var/opt/secondlife/SecondLife-linux-02012006/bin/secondlife-bin)
==5902== by 0x9825412: LLViewerImageList::updateImages(float) (in /var/opt/secondlife/SecondLife-linux-02012006/bin/secondlife-bin)
==5902== by 0x9DA6865: idle() (in /var/opt/secondlife/SecondLife-linux-02012006/bin/secondlife-bin)
2006-02-02T12:57:27Z INFO: remove_marker_file()

This bug has completely gone away in 1.8.3(9). Hooray!