bug#37215: [PATCH] vc-cvs-ignore writes absolute filenames and duplicate strings

[Wolfgang Scherer]

Am 21.02.20 um 10:31 schrieb Eli Zaretskii:
>> Cc: larsi@gnus.org, 37215@debbugs.gnu.org, dgutov@yandex.ru
>> From: Wolfgang Scherer <Wolfgang.Scherer@gmx.de>
>> Date: Thu, 20 Feb 2020 00:06:11 +0100
>> (lisp/vc/pcvs.el) Call ‘vc-cvs-append-to-ignore’ with SORT argument.
> Please quote 'like this' in log messages, and try to avoid non-ASCII
> characters there (they are generally only necessary in people's
> names).
OK
>> +Patterns follow glob(7) syntax.  Special characters \"?*[\\\" are
>> +escaped with a backslash."
> I'd say "should be escaped" here, since this is a requirement for the
> argument passed to this function.
Here is that amgbuity again ;-). It is only required, if the user
wants a special character to match literally. It's perfectly fine to
specify  *.pyc  as a pattern. I have phrased it like that.
> Also, I'd mention that FILE can be a pattern, otherwise the reference
> to patterns might come as a surprise to the reader.
I emphasized more, that the basename of the FILE argument (not the
entire FILE) is in fact a CVS ignore pattern.
>> +to hear about anymore.  If SORT is non-nil, sort the ines of the
>> +ignore file."                                        ^^^^
> Typo: should be "lines".
Right.
>> +    (goto-char (point-min))
>> +    (save-match-data
>> +      (unless (re-search-forward (concat "^" (regexp-quote str) "$") nil t)
>> +        (goto-char (point-max))
> You could use non-nil, non-t 3rd argument of re-search-forward, in
> which case the following goto-char would be redundant, right?
Right, I just left the  goto-char  in there, because it makes it
obvious what is going on. Switching to the side-effect optimization ...

0001-Do-not-write-absolute-filenames-and-duplicate-string.patch
Description: Text Data