Stefan Sperling <s...@elego.de> writes: > I'll investigate a bit further, but if my testing shows that this > change is reponsible I am going to revert it ASAP to prevent the > damage from spreading. > > 'svn diff' coredumps with: > > (gdb) > #0 0x000011bd8dc946bd in svn_eol__find_eol_start (buf=0x11bd84838ffa "", > len=12309) at subversion/libsvn_subr/eol.c:56 > 56 apr_uintptr_t chunk = *(const apr_uintptr_t *)buf; > (gdb) > Bottom (i.e., innermost) frame selected; you cannot go down. > (gdb) p buf > $1 = 0x11bd84838ffa "" > (gdb) p chunk > $2 = 0 > (gdb) p * buf > $3 = 0 '\0'
I see the same: ==5285== Invalid read of size 8 ==5285== at 0x5A3EF29: svn_eol__find_eol_start (eol.c:56) ==5285== by 0x5A6D41B: translate_chunk (subst.c:1144) ==5285== by 0x5A6DB9D: translated_stream_write (subst.c:1323) ==5285== by 0x5A657C9: svn_stream_write (stream.c:161) ==5285== by 0x5A662E2: svn_stream_copy3 (stream.c:501) ==5285== by 0x5A6EC6C: svn_subst_copy_and_translate4 (subst.c:1793) ==5285== by 0x512028C: svn_wc__internal_translated_file (translate.c:226) ==5285== by 0x50F0F0B: svn_wc__diff_base_working_diff (diff_editor.c:526) ==5285== by 0x50F8064: diff_status_callback (diff_local.c:362) ==5285== by 0x511B731: send_status_structure (status.c:930) ==5285== by 0x511BC7A: one_child_status (status.c:1206) ==5285== by 0x511C4F4: get_dir_status (status.c:1441) ==5285== Address 0x40feffb is not stack'd, malloc'd or (recently) free'd -- Certified & Supported Apache Subversion Downloads: http://www.wandisco.com/subversion/download
