Re: Patch: rename to `insight'

[Eli Zaretskii <eliz at delorie dot com>]

> Date: Wed, 29 Nov 2000 14:59:35 -0500
> From: Christopher Faylor <cgf@redhat.com>
> >
> >I assume you're talking about Windows.  I thought gdb required Cygwin
> >to run on Windows, in which case you will have a shell.  Or is there a
> >real native gdb?
> 
> I think that the mingw project has a native version of gdb but I think
> that it still uses '/' rather than '\' for its paths.

If so, that's a bug in the MinGW port, which I hope someone, some day,
will fix.  IMHO it is not a good idea to rely on unfixed bugs in GDB.

In any case, I think it isn't nice to require users to install a shell
if they can do without it.

> I don't believe that it works with insight, though.

Probably because no one bothered to port it.  Someone might do that in
the future, though, so if it's possible, I think we shouldn't add
obstacles on that someone's way.

> AFAICT, this shell script would work as well on Windows/Cygwin as it
> does on linux.

Cygwin forces the Posix environment on all the tools, but other
(non-Cygwin) toolchains support native syntax and conventions as well.

My point was that, if we use a program, the gory details of these
problems are concealed in the library, whereas in a shell script you
need to test for them explicitly.  Consider the mess introduced into
libtool and the configure scripts in order to detect Cygwin, MinGW,
and whatnot, and then DTRT with backslashes and drive letters brain
damage.

While libtool and configure are auto-generated by tools that are
explicitly tailored to shell scripts, and thus their maintainers had
little choice when adapting to non-Posix platforms, the script written
by Tom is not produced by those automated tools.  Thus my suggestion
to throw together a short program instead of a shell script.