Hi Helmut: Am 11.09.17 18:54 schrieb(en) Helmut Jarausch:
thanks for the suggestions but valgrinds output doesn't say much to me: =19859= Memcheck, a memory error detector =19859= Copyright (C) 2002-2015, and GNU GPL'd, by Julian Seward et al. =19859= Using Valgrind-3.12.0 and LibVEX; rerun with -h for copyright info =19859= Command: src/balsa =19859= Parent PID: 2090 =19859= =19859= Thread 7: =19859= Syscall param write(buf) points to uninitialised byte(s) =19859= at 0xBBDEFF4: write (in /lib64/libpthread-2.26.so) =19859= by 0xB4EB214: write_to_temp_file (in /usr/lib64/libglib-2.0.so.0.5200.3) =19859= by 0x20BBD96F: ??? =19859= Address 0x1caaa254 is 512,884 bytes inside a block of size 524,288 alloc'd =19859= at 0x4C2DE8F: malloc (vg_replace_malloc.c:298) =19859= by 0x4C30254: realloc (vg_replace_malloc.c:785) =19859= by 0xB504C9F: g_realloc (in /usr/lib64/libglib-2.0.so.0.5200.3) =19859= by 0x7FFFF: ??? =19859= by 0xB4D0DB8: g_array_maybe_expand (in /usr/lib64/libglib-2.0.so.0.5200.3)
This one looks somewhat harmless afaict, and at least the call stack is intact…
=19859= Thread 1: =19859= Invalid read of size 8 =19859= at 0xB4FE8A7: g_main_context_prepare (in /usr/lib64/libglib-2.0.so.0.5200.3) =19859= by 0xF00000000: ??? =19859= by 0x19FB350F: ??? =19859= by 0x1CDEC54F: ??? =19859= by 0x1FA05DFB: ??? =19859= by 0xCAE39A4EB8AF4FFF: ??? =19859= by 0x27: ??? =19859= by 0x19DF8B8F: ??? =19859= by 0x1FA05E8F: ??? =19859= by 0x1D1B0DEF: ??? =19859= by 0xFFEFFE9A3: ??? =19859= Address 0x22552880 is not stack'd, malloc'd or (recently) free'd
…but this one doesn't, of course. And again the stack looks completely broken. Did you try valgrind's suggestions, i.e. enlarge the stack size and add “-v --track-origins=yes” to the command line? Does the crash again happen when you try to launch the message composer, or before you click “Send”? In the latter case, what happens if you cancel the send operation instead? Any crash or serious errors in valgrind? I must admit that I'm a little lost here. I tried the two different Balsa configurations you posted earlier, both on a Debian 32-bit (jessie) vm and on a native 64-bit Ubuntu 16.04, and could not reproduce *any* crash (but the linker issue with balsa-ab with the second one…). I will try to install a 64-bit Gentoo vm and check Balsa there. This may take some time, though, as I'm not familiar with it. Cheers, Albrecht.
Attachment:
pgpc6S00iMvRl.pgp
Description: PGP signature