Re: CVS broken
Adrian Aichner <adrian(a)xemacs.org> writes:
>>>>> "Dan" == Dan Holmsand
<dan(a)eyebee.com> writes:
Dan> "Stephen J. Turnbull" <turnbull(a)sk.tsukuba.ac.jp> writes:
>> I hate computers---it decided to belch out the diff just as I sent
>> that last. There is Dan Holmsand change I don't fully understand in
>> fileio.c, in Ffile_truename, but it shouldn't be semantically
>> different under Unix. Is there a reason not to use
>>
>> p = (Extbyte *) memchr (p + 1, DIRECTORY_SEP, elen - (p + 1 - path))
>>
>> instead of the explicit loop?
>>
Dan> Yes. On windows both '/' and '\' should be interpreted as
directory
Dan> separators.
Wouldn't it be better to obey the value of
`directory-sep-char' (buffer: *wide reply to Dan Holmsand*, mode: Message)
No (incidentally, that's the same as DIRECTORY_SEP on windows).
Something like C:/temp is a valid filename even if directory-sep-char
is ?\\. Filenames like c:/temp\foo.txt are also valid, of course.
However, file-truename does obey directory-sep-char, in that it uses
the 'correct' separator in the result.
/dan
--
Dan Ola Holmsand
IB
dan(a)eyebee.com