Re: canonicalize_file_name simplification?

[Rik]

On 12/03/2013 01:54 PM, address@hidden wrote:
> Message: 7
> Date: Tue, 3 Dec 2013 12:23:55 -0800 (PST)
> From: PhilipNienhuis <address@hidden>
> To: address@hidden
> Subject: Re: canonicalize_file_name simplification?
> Message-ID: <address@hidden>
> Content-Type: text/plain; charset=us-ascii
>
> Rik-4 wrote
>> > 12/3/13
>> > 
>> > John,
>> > 
>> > In liboctave/file-ops.cc there is a 4-way #ifdef test to determine whether
>> > there is some sort of function we can use for canonicalize_file_name.  The
>> > end result is that I am seeing very different behavior between Linux and
>> > __WIN32__ platforms.  Given that we have adopted gnulib and are using the
>> > canonicalize module, can't we just call gnulib::canonicalize_file_name and
>> > let it sort out which function to call?
> Would the canonicalize_file_name issues in bug #36677 (notably for not
> (-yet) existing files on Win32) interfere here?
The issue stated in the title for bug #36677 is now fixed and there is no
inconsistency between platforms.  On the other hand, you were looking for
additional behavior when the input filename and path does not exist.  You
might want to re-title the bug report to reflect that issue which is that
make_absolute_filename needs to be extended to deal with drive letters on
windows platforms.

--Rik