Re: [gdb/libiberty] Improve support for cross debugging shared libraries with DOS style pathnames (from Unix hosts)

[Eli Zaretskii <eliz at gnu dot org>]

> From: Pedro Alves <pedro@codesourcery.com>
> Date: Fri, 23 Apr 2010 11:46:22 +0100
> Cc: gcc-patches@gcc.gnu.org
> 
> Oh.  I really meant a "path", as in:
> 
>  "A path, the general form of a filename or of a directory name, specifies a unique
>  location in a file system."
> 
>  (From <http://en.wikipedia.org/wiki/Path_%28computing%29>.)
> 
> I suppose that's why we have IS_ABSOLUTE_PATH, and not IS_ABSOLUTE_FILE_NAME,
> and such.
> 
> IMO, it's the use of singular path or "search path" to mean
> "list of directories" that's overloaded.
> "Set assumed file system kind for target reported file names"
> kind of sounds like it is only talking about the filename "/dir1/dir2/files1"
> (or directory and filenames, if you think of a directory a file as well).
> Maybe I could just get rid of the mention of paths/file names, and go with:
> 
>   "Set the assumed file system kind of the target."
> 
> Anyway, I'll change it to what you think is clearer.

Not only is it clearer, it's in GNU coding standards, see my other
mail.

>   c:/foo/bar.dll ==> /path/to/sysroot/foo/bar.dll
> "
> 
> I'll be happy to clarify this in the docs, if you think it's useful.

No, it's clear enough.  I just wondered why you didn't remove the
drive letter right away.  I guess that's common practice in some
quarters to mount DOS-ish filesystems that way.  I don't really mind
to support that if people are using it.

> I suppose that'll be ... in `i386_go32_init_abi'.

Yep.

> (Why are DJGPP specific things in the generic i386-tdep.c, BTW?)

History, I think.  The number of architecture specific bits is quite
small.

> >   @cindex DOS file-name semantics if file names.
> 
> Will do.  I'll assume you meant s/if/of/.

Of course.

Thanks.