Date: Wed, 1 Sep 1999 12:19:37 +0300 (IDT) From: Eli Zaretskii X-Sender: eliz AT is To: "Mark E." cc: djgpp-workers AT delorie DOT com Subject: Re: Bug in collect2.c In-Reply-To: <199908311420.OAA79880@out2.ibm.net> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Reply-To: djgpp-workers AT delorie DOT com X-Mailing-List: djgpp-workers AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk On Tue, 31 Aug 1999, Mark E. wrote: > I think first set of slashes and backslashes should be collapsed > following a drive letter and colon: c:/\ld.exe c:\/ld.exe c:/\/\ld.exe. ``First set''? I would rather think ``last set'', that is, the one immediately preceding the basename. If we want to fix problems with pathnames constructed by programs that don't treat \ as /, then they usually happen when a file name is appended to a leading directory. > \\unc\dir\file is certainly a > valid UNC, but should c:\\unc\dir\file be considered valid too? I don't know; it might. See the thread on c.o.m.d., where some people reported that \\foo and c:\\foo causes the same timeout-like delay. Until now, the policy in _putpath was to leave backslashes unaltered. Since the problem that started this thread was not with two consecutive backslashes, perhaps we shouldn't fix what isn't broken, and leave consecutive backslashes alone.