On Monday, September 17, 2012 3:29:56 PM UTC-4, Christian Brabandt wrote:
> Hi Donald!
>
>
>
> On Sa, 15 Sep 2012, Donald Allen wrote:
>
>
>
> > I am running vim 7.3.659 on a 64-bit Arch Linux system. I am seeing
>
> > what I believe to be a bug in the filename-completion logic as used in
>
> > the :buffer command. To reproduce, do the following:
>
> >
>
> > 1. mkdir /tmp/foo
>
> > 2. cd /tmp/foo
>
> > 3. Somehow, create two files in /tmp/foo -- foo.c and bar.c.
>
> > 4. Start vim. Enter ':e bar' followed by <tab>. It will complete to
>
> > 'bar.c'. Hit <enter> to load the file. Do the same for foo.c. Note
>
> > that after typing ':e foo' and then <tab>, it will complete to
>
> > 'foo.c'.
>
> > 5. You will have foo.c displayed in the vim window. If you enter ':b
>
> > ba' followed by <tab>, it will complete to 'bar.c'. Hit <enter>. Now
>
> > the window displays bar.c. Now type ':b fo' followed by <tab>; it
>
> > will not complete. But if you hit <enter> at that point, the window
>
> > will switch correctly to foo.c. I think the logic is getting confused
>
> > by the directory name and the filename both being 'foo'. If you move
>
> > the directory to /tmp/baz, the problem will not occur.
>
>
>
> Hm, that might be a bug. The problem is, that :b fo uses a different
>
> algorithm for determining the buffer, than the completion function.
>
>
>
> That is not the bug, so. Anyhow, I think ExpandBufnames can try to first
>
> match files explicitly, before falling back to the old behaviour. But I
>
> am not sure, this is useful.
>
> Something like this:
>
>
>
> diff --git a/src/buffer.c b/src/buffer.c
>
> --- a/src/buffer.c
>
> +++ b/src/buffer.c
>
> @@ -2236,24 +2236,29 @@
>
> /* Make a copy of "pat" and change "^" to "\(^\|[\/]\)". */
>
> if (*pat == '^')
>
> {
>
> - patc = alloc((unsigned)STRLEN(pat) + 11);
>
> + patc = alloc((unsigned)STRLEN(pat) + 29);
>
> if (patc == NULL)
>
> return FAIL;
>
> STRCPY(patc, "\\(^\\|[\\/]\\)");
>
> STRCPY(patc + 11, pat + 1);
>
> + /* Make sure, patc only matches files for 1st attempt */
>
> + STRCPY(patc + STRLEN(patc), "[^\\/]*\\([\\/]\\)\\@!$");
>
> }
>
> else
>
> patc = pat;
>
>
>
> /*
>
> - * attempt == 0: try match with '\<', match at start of word
>
> + * attempt == 0: try match with '\<', match at start of word and
> force a match of a file
>
> * attempt == 1: try match without '\<', match anywhere
>
> */
>
> for (attempt = 0; attempt <= 1; ++attempt)
>
> {
>
> if (attempt > 0 && patc == pat)
>
> break; /* there was no anchor, no need to try again */
>
> - prog = vim_regcomp(patc + attempt * 11, RE_MAGIC);
>
> + else if (attempt > 0)
>
> + STRCPY(patc, pat + 1);
>
> +
>
> + prog = vim_regcomp(patc, RE_MAGIC);
>
> if (prog == NULL)
>
> {
>
> if (patc != pat)
>
>
>
>
>
> regards,
>
> Christian
>
> --
>
> Das Menschenleben ist seltsam eingerichtet: Nach den Jahren der Last
>
> hat man die Last der Jahre.
>
> -- Johann Wolfgang von Goeth
Thanks for the response and the patch. I will give it a try and report back.
/Don Allen
--
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
