On 8/6/07, Bram Moolenaar <[EMAIL PROTECTED]> wrote:
> Patch 7.1.055
> Problem: Using strcpy() with arguments that overlap.
> Solution: Use mch_memmove() instead.
> *** ../vim-7.1.054/src/termlib.c Thu May 10 20:20:59 2007
> --- src/termlib.c Sun Aug 5 21:52:41 2007
> ***************
> *** 191,197 ****
> lbuf[0] == '\t' &&
> lbuf[1] == ':')
> {
> ! strcpy(lbuf, lbuf+2);
> llen -= 2;
> }
> if (lbuf[llen-2] == '\\') /* and continuations */
> --- 191,197 ----
> lbuf[0] == '\t' &&
> lbuf[1] == ':')
> {
> ! mch_memmove(lbuf, lbuf + 2, strlen(lbuf + 2) + 1);
> llen -= 2;
> }
> if (lbuf[llen-2] == '\\') /* and continuations */
I've noticed that STRLEN() wrapper is used everywhere in
this patch 55 (and in vim in general), except in the last chunk
of the patch, where standard C strlen() is used. I assume it
should be STRLEN().
Also, regarding replacing STRCPY(p, p + 2); with
mch_memmove(p, p + 2, STRLEN(p) - 1);
I assume it's impossible to have p point to a buffer containing
"\0\0\0" for example, or else it would crash (STRLEN(p) - 1
being a negative number).
I think mch_memmove(p, p + 2, STRLEN(p + 2) + 1); is thus closer to
the original behavior.
Also, if p is not an empty string, mch_memmove(p, p + 2, STRLEN(p + 2) + 1);
seems slightly more efficient than mch_memmove(p, p + 2, STRLEN(p) - 1);
since it saves 2 tests and 2 loop iterations inside STRLEN for only an
additional +2 constant addition (better for pipeline)
Cheers
-- Dominique
--~--~---------~--~----~------------~-------~--~----~
You received this message from the "vim_dev" maillist.
For more information, visit http://www.vim.org/maillist.php
-~----------~----~----~----~------~----~------~--~---
