This is the mail archive of the
gdb@sources.redhat.com
mailing list for the GDB project.
Re: I have found, why I get the exec file does not match error ...
- To: reinhard dot jessich at telering dot at
- Subject: Re: I have found, why I get the exec file does not match error ...
- From: egor duda <deo at logos-m dot ru>
- Date: Mon, 18 Jun 2001 21:12:24 +0400
- CC: gdb at sources dot redhat dot com
- Organization: deo
- References: <01061722341402.07777@localhost.localdomain><01061816235800.02941@localhost.localdomain><178357917037.20010618190950@logos-m.ru><01061818595600.03333@localhost.localdomain>
- Reply-To: egor duda <deo at logos-m dot ru>
Hi!
Monday, 18 June, 2001 reinhard.jessich@telering.at reinhard.jessich@telering.at wrote:
ED>> no, i'm starting it with 'gdb --core=xxx.core'
rjta> This is possible, but then I will not have the symbols!
rjta> Have you symbols?
rjta> If yes, I would be interested where this symbols are loaded.
they're loaded when core is loaded. core contains a path to a dumped
process' executable and dlls, so it can load symbols from them. of
course, it works if you debug your dump on the same host where it's
created, or, at least, they have same files in the same paths.
rjta> Can you try to start it with gdb xxx.exe xxx.core. I would be intereset, if you
rjta> get then the same warning.
yes. i suppose, however, that core_matches_executable_p () logic is
somewhat flawed. I don't see why bfd vectors for executable and core
should necessarily be the same. if executable is in pei format, but
core is elf, i don't feel it's necessarily wrong. i don't know if it's
possible to make this check more sophisticated, yet foolproof, however.
Egor. mailto:deo@logos-m.ru ICQ 5165414 FidoNet 2:5020/496.19