The exception dump is absolutely NOT useful (unless the person in question sends a copy of his current memory-dump - and I mean completely - also :) ).

Luckily (Ah, I did explain this quite some times in the past, maybe I should put it into a textfile for future reference ? :) ) there is a way to get very detailed info:

1. Install segtracker (call "c:segtracker >nil:" shortly after setpatch in your startup-sequence, segtracker is if I remember right available on the Developer CD, if you cannot find it, ask me, I could send it, it is only some hundred bytes sized)
2. Install sushi or sashimi (should be on Aminet). Copy the program to c:
3. Type in a shell setenv "100" >envarc:powerpc/seginfo
(you only need to do this once)
4. Reboot
5. Now start sushi like this: sushi >ram:test.txt
6. Now start the crashy program
7. Abort sushi with CTRL-C after the crash
8. Send test.txt to the author of the program.

9.

The author of the program NEEDS to have kept a copy of the ELF-File from which he generated the exe, and then do

pobjdump -d --start-address=0x... name-of-the-elf-file

the 0x... data will be in test.txt. Please note, that only the first entry will make sense (WarpOS assumes the program is PowerOpen, but with gcc-WarpOS it is internally System V). You can get a complete stacktrace by "feeding" the R1 register of the crash (or LR, if r1 contains nonsense) to the checkstack tool (should be on Aminet).
(checkstack needs to be done by the person who got the crash, of course... the pobjdump by the program author, though).

If I remember right --source can be added to pobjdump, if the program was compiled with debug-output. Then you get the exact source-line of the crash. If not you just get the name of the crashed function AND the ASM output at that place.

Note that everything below "8." is gcc-WarpOS specific. Programs compiled with StormC or vbcc (using PowerOpen) do it different (simpler :) ).

Hope this helps. If there are questions left, send me an email.

(Phew, under OS 4 this is so much easier to get a stacktrace... the crash-requester already includes the stacktrace, no need for extra tools anymore :) ).

Steffen

steffen: yes i know, i emailed rich but i havent heard anything yet.

i could ofcourse give a full sashimi/segtrack if someone asked me to give one,
i just posted that "wos exception" to show how horrible it is :) (makes no sense to anyone)..


anyway thanx for the info..

cheers

_________________
See my blog and collection website! . https://www.blog.amigaguru.com

ok since rich is not answering my emails i might aswell post it here..

for now..

though this output might be of no use either..

--

sushi installed (CTRL-C or BREAK to remove)
Ignore this if you are running Enforcer RAWIO 37.x or higher
Enforcer 2.8b (or 2.6f megastack) could not be found or patched.
If you start 2.8b or 2.6f now, stop and restart sushi to patch Enforcer.

PowerPC Instruction access exception
Page fault
The next instruction to be fetched is located at FFFFFFFC
Task name: 'uae_PPC10' Task address: 70F5E630 Task ID: 110

SRR0: FFFFFFFC SRR1: 4000F070 MSR: 00003040 HID0: 0040C000
PVR: 00071201 DAR: 00000000 DSISR: 00000000 SDR1: 7080003F
DEC: 00002E8E TBU: 00000001 TBL: 5D2D56C6 XER: 00000000
CR: 80000000 FPSCR: 00000000 LR: 72FE45B0 CTR: 00000000

R0-R3: FFFFFFFF 735DBD10 70F5EB08 00000000 IBAT0: 00001FFF 00000022
R4-R7: FFFFFFFF 00000000 00000000 80000000 IBAT1: 00000000 00000000
R8-R11: 00000025 FFFFFFFF 00000000 00000000 IBAT2: 00000000 00000000
R12-R15: 00000008 735DBE80 706B4418 706B890C IBAT3: 00000000 00000000
R16-R19: 735DBDC0 0000004C 00000000 00000000 DBAT0: 00001FFF 0000002A
R20-R23: 00000000 00000000 7FE980DA 000003ED DBAT1: 00000000 00000000
R24-R27: FFFFFF89 7FE98096 00000007 00000000 DBAT2: 00000000 00000000
R28-R31: 00000000 00000000 FFFFFFFF 00000000 DBAT3: 00000000 00000000

F0-F3: 0 2.5039 E-312 1.5153 E 244 2.1607 E 244
F4-F7: 1.5071 E 244 9.5018 E-315 QNAN 1.5070 E 244
F8-F11: 1.5079 E 244 0 0 0
F12-F15: 0 0 0 0
F16-F19: 0 0 0 0
F20-F23: 0 0 0 0
F24-F27: 0 0 0 0
F28-F31: 0 0 0 0

Warp-OS 2000 by Sam Jordan

Segment statistics

706B4410 : "LIBS:powerpc.library" Hunk 0 Offset 00012880
706A84A0 : "LIBS:powerpc.library" Hunk 0 Offset 00006910
sushi removed

_________________
See my blog and collection website! . https://www.blog.amigaguru.com

Quote:

This is great... Looking forward to the AOS4 version maybe a contrib on the CD I ask.. ?

An AOS4 version is planned.

Problem is - I'd guess - the seginfo-stuff of WOS cannot cope very well with System V Code, which gcc-WOS internally uses. Best "go" is to use checkstack and memowner on R1 (and possibly memowner on LR) to get more info. memowner is included with StormC, for example, but I THINK it is also on the developer CD. checkstack (tool I wrote, to run through a System V stackframe) is on Aminet I think...

Steffen

@ evilrich

"A new release of my unofficial, experimental version of UAE for Linux, Unix, OS X, BeOS and AmigaOS is available for download."

Do you still intend to provide binaries for Linux/PPC glibc 2.2? (I know you had problems with your glibc 2.2 setup).