Add source context where available about gef HOT 16 CLOSED

That's a pretty good idea. Will add that.

Thanks!

from gef.

It should be working as expected. Let me know if you find bugs.

from gef.

Confirmed works on x86, ARM, PowerPC for gdb > 7.9 (for gdb.find_pc_line())

from gef.

By >7.9 you mean it only works on 7.10? I re-enabled the exception printing and I get:

[!] in `context_source, exception raised: 'utf-8' codec can't decode byte 0xa9 in position 96: invalid start byte

ben@shitweasel:~/src/jxrlib$ gdb -version
GNU gdb (Ubuntu 7.9-1ubuntu1) 7.9

from gef.

Which version of python are you using inside gdb ?

from gef.

I don't know how to check that. System python is 2.7 but I thought gdb always links against 3 now? I'll google tomorrow :P

from gef.

Open a python shell in gdb ("pi" command), then

import sys
sys.version

from gef.

@bnagy $ gdb -nx -ex 'python print (sys.version)' -ex quit

from gef.

'3.4.3 (default, Mar 26 2015, 22:07:14) \n[GCC 4.9.2]'

from gef.

I've setup an Ubuntu 14.04 with gdb 7.7 and python 3.4, could get your error triggered again.

Is it happening against all your bins ? Or just one ?

from gef.

It's just one, by the looks - AND the source file happens to have a non-utf-8 character in the source file, so that sounds like a good chance to be the problem. I tried a quick guess from @bsmt using return buf.decode("utf-8", "ignore") in read_string but that didn't work - maybe I missed something or that's the wrong place to try and ignore bad UTF-8.....

from gef.

Do you have a sample somewhere I can download so I try to reproduce locally ?

from gef.

Can't really, with this one... you should be able to put any non utf-8 char in the source though like from http://www.cl.cam.ac.uk/~mgk25/ucs/examples/UTF-8-test.txt

For now, I think I can run the codebase through iconv, which solves my immediate issue as per http://superuser.com/a/49147 (which is a possible workaround for anyone else with the issue, but hopefully it's easy to fix ;)

from gef.

Have you tried changing the charset (https://sourceware.org/gdb/onlinedocs/gdb/Character-Sets.html#Character-Sets) ? (set target-charset in gdb)

from gef.

I tried set target-charset WINDOWS-1252 with no luck. Fixing the source files via iconv still works. If you can't work out a way to clean / change the encoding in the plugin then it's probably OK to close this issue, or just open a new one for the encoding problem (since the enhancement is complete)

Thanks again for the tool :)

PS: Inplace fixing didn't seem to work for me. This script does.

#! /bin/sh

find . -name '*.c' -exec iconv --verbose -f windows-1252 -t utf-8 -o {}.fix {} \;
find . -name "*.c" -delete
find . -name "*.c.fix" -exec rename 's/\.fix$//' {} \;

from gef.

I haven't been able to find a clean so I guess I will do as you suggest and close the ticket keeping in mind your little script in case this issue happens again in the future.

Thanks for the feedbacks, always appreciated :)

from gef.