Dominique wrote:
> When running test32 with vim-7.4.22, I see this valgrind error:
>
> ==17589== Memcheck, a memory error detector
> ==17589== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al.
> ==17589== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info
> ==17589== Command: ../vim -u unix.vim -U NONE -c set\ maxmem=256
> --noplugin -s dotest.in test32.in
> ==17589== Parent PID: 16711
> ==17589==
> ==17589== Invalid read of size 1
> ==17589== at 0x4BCE96: utf_head_off (mbyte.c:3434)
> ==17589== by 0x424339: ins_complete (edit.c:5191)
> ==17589== by 0x42A7C1: edit (edit.c:1410)
> ==17589== by 0x4C10F7: invoke_edit (normal.c:9232)
> ==17589== by 0x4C3721: nv_edit (normal.c:9205)
> ==17589== by 0x4C9014: normal_cmd (normal.c:1200)
> ==17589== by 0x576AE9: main_loop (main.c:1329)
> ==17589== by 0x579C1B: main (main.c:1020)
> ==17589== Address 0xc3daa4f is 1 bytes before a block of size 8 alloc'd
> ==17589== at 0x4C2B3F8: malloc (in
> /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==17589== by 0x4B5DE6: lalloc (misc2.c:929)
> ==17589== by 0x4B5E8B: alloc_check (misc2.c:862)
> ==17589== by 0x4A975E: ins_str (misc1.c:2347)
> ==17589== by 0x42703D: insertchar (edit.c:6043)
> ==17589== by 0x4279DB: insert_special (edit.c:5806)
> ==17589== by 0x42AC3C: edit (edit.c:1491)
> ==17589== by 0x4C10F7: invoke_edit (normal.c:9232)
> ==17589== by 0x4C3721: nv_edit (normal.c:9205)
> ==17589== by 0x4C9014: normal_cmd (normal.c:1200)
> ==17589== by 0x576AE9: main_loop (main.c:1329)
> ==17589== by 0x579C1B: main (main.c:1020)
>
>
> Line edit.c:5191 was recently modified in this patch:
>
> changeset: 5332:d5eb32dc231c
> tag: v7-4-019
> user: Bram Moolenaar <[email protected]>
> date: Thu Sep 05 13:50:53 2013 +0200
> files: src/edit.c src/version.c
> description:
> updated for version 7.4.019
> Problem: MS-Windows: File name completion doesn't work properly with
> Chinese characters. (Yue Wu)
> Solution: Take care of multi-byte characters when looking for the start of
> the file name. (Ken Takata)
>
> $ hg annotate edit.c:
>
> 5184 vimboss 7: else if (ctrl_x_mode == CTRL_X_FILES)
> 5185 vimboss 7: {
> 5186 bram 5332: char_u *p = line + startcol;
> 5187 bram 5332:
> 5188 bram 5332: /* Go back to just before the first filename
> character. */
> 5189 bram 5332: mb_ptr_back(line, p);
> 5190 bram 5332: while (vim_isfilec(PTR2CHAR(p)) && p >= line)
> !5191 bram 5332: mb_ptr_back(line, p);
> 5192 bram 5332: startcol = p - line;
> 5193 bram 5332:
> 5194 vimboss 449: compl_col += ++startcol;
> 5195 vimboss 449: compl_length = (int)curs_col - startcol;
> 5196 vimboss 449: compl_pattern = addstar(line + compl_col,
> compl_length,
> 5197 vimboss 449: EXPAND_FILES);
> 5198 vimboss 449: if (compl_pattern == NULL)
> 5199 vimboss 7: return FAIL;
> 5200 vimboss 7: }
>
> Attached patch fixed it but please verify it.
Thanks for the patch!
Theoretically it might still be invalid to have a pointer just before a
string. I never encountered a machine with that problem, but it's nicer
to have starcol not become negative and then increment it. Thus I'll
change it slightly:
mb_ptr_back(line, p);
while (p > line && vim_isfilec(PTR2CHAR(p)))
mb_ptr_back(line, p);
startcol = (int)(p - line) + 1;
if (p == line && vim_isfilec(PTR2CHAR(p)))
startcol = 0;
compl_col += startcol;
--
hundred-and-one symptoms of being an internet addict:
187. You promise yourself that you'll only stay online for another
15 minutes...at least once every hour.
/// Bram Moolenaar -- [email protected] -- http://www.Moolenaar.net \\\
/// sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\ an exciting new programming language -- http://www.Zimbu.org ///
\\\ help me help AIDS victims -- http://ICCF-Holland.org ///
--
--
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php
---
You received this message because you are subscribed to the Google Groups
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
