Re: Emacs vista build failures

[David Kastrup <dak@gnu.org>]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: David Kastrup <dak@gnu.org>
>> Date: Tue, 22 Jul 2008 16:04:03 +0200
>> Cc: Johannes Weiner <hannes@saeurebad.de>, emacs-devel@gnu.org
>> 
>> REM Get current directory
>> set targetdir=%CD%
>> 
>> REM %targetdir% has to be postprocessed since if it is a root directory,
>> REM it will end in a backslash which escapes a double quote when calling
>> REM Java.  So we append a single dot in that case.
>> for %%l in (%targetdir%) do if "%%~pnl" == "\" (set targetdir=%targetdir%.)
>> 
>> call ant.bat -Dinstaller.path="%targetdir%"
>> 
>> 
>> Can you imagine how many mandays get wasted on utterly appalling
>> workarounds like that?  And of course, this just works on one version of
>> cmd.exe, and might break on another.
>> 
>> Bourne shell programming is _much_ more consistent, regular, and simple.
>
> Again, you are approaching a non-Posix platforms with Posix-centric
> perspective.  Shell programming is a Posix idea.  If you want a
> program, write it in a programming language, not in shell.

So what programming languages does Windows come with?  Or are we not
actually talking about Windows after all?  "Windows is great, because
you can get something else to work around its shortcomings."?

>> This list certainly is not the place to discuss the presence or
>> absence of merits in Windows.  But if we get a bit more
>> Emacs-specific and you take a look at conditional code being used
>> when under all sort of UNIX systems and code being used when under
>> w32, then take a good look at what code is more complex and awkward.
>
> I already explained why: the original code was designed with Posix
> functionality in mind, that's why it doesn't port easily to anything
> else.

Whatever.  It is clear that you will not be fazed from your claims.  So
it is rather pointless to continue this charade.

-- 
David Kastrup