Heh, Yeah, I agree with you both! I think it would
be a lot cleaner and easier to maintain if it was
all rewritten in high-quality C++, but I haven't
had time to make a patch yet. ;)
- Rob
--
Robert Boehne Software Engineer
Ricardo Software Chicago Technical Center
TEL: (630)78
Albert Chin wrote:
> Isn't the use of `$echo $CC` gross?
Isn't the whole script just a bit gross anyway? ;-)
___
Libtool mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/libtool
On Tue, Apr 02, 2002 at 10:41:15AM -0600, Robert Boehne wrote:
> There was a bit missing in my patch yesterday, which the tagdemo
> tests caught right away. The same logic to check for spaces needs
> to be in the tag matching code all the way through. With this patch
> I'm getting all tests pass
Hello!
There was a bit missing in my patch yesterday, which the tagdemo
tests caught right away. The same logic to check for spaces needs
to be in the tag matching code all the way through. With this patch
I'm getting all tests passed under Tru64 and Linux, and no more
than two failures (namely
I forgot the patch again!
BTW: I found this is a lot better than trying to
strip off a leading space. Also, I found that it
was necessary to check both cases, with a space prepended,
and without.
Thanks!
--
Robert Boehne Software Engineer
Ricardo Software Chicago Technical Cent
Bruce:
Now that I look at it, it may be better to remove
that initial space from $base_compile.
Any thoughts on that?
--
Robert Boehne Software Engineer
Ricardo Software Chicago Technical Center
TEL: (630)789-0003 x. 238
FAX: (630)789-0127
email: [EMAIL PROTECTED]
_
Robert Boehne wrote:
>
> Bruce,
>
> Ok, I found the problem.
THANK YOU! Actually, the way the code is written, I think
you can rest assured that the first character of base_compile
will always be a space. All the text insertions look like
this:
base_compile="$base_compile $lastarg"
Of
Bruce,
Ok, I found the problem. Before your recent patch, either
$base_compile never had a space as it's first character, or
it was stripped out somewhere (maybe by accident?).
Regardless, the problem was that in the check for a valid
tag, $base_compile=" $CC ..." was not checked for, and THAT
w
Bruce,
Ok, I found a big problem here, it seems that
currently Libtool will ignore all tags, and use
CC in every case. That's bad. But along the
reasoning that spurned your original patch, I don't
want to un-do it. I will work on a "proper" fix
if I can figure out how Libtool ever managed to p
Bruce:
I found a few things to improve on. These are pretty
minor changes that fix an obvious problem, so I'm
going to take the liberty of checking them in asap.
Thanks!
2002-03-14 Robert Boehne <[EMAIL PROTECTED]>
ltmain.in: Touch-up to make testsuite pass, and default tagname
Robert Boehne wrote:
>
> Bruce:
>
> ... I'd say "go ahead, if it breaks I'll let you know." ;)
> The -Xcompiler thingy is *very* necessary though, so
> make try not to break it.
OK. I made the try. :-)
If you would try to poke a hole in:
http://autogen.sourceforge.net/data/ltmain.in
I'd a
Bruce:
I don't have a problem with you cleaning up things in
CVS head, I'd say "go ahead, if it breaks I'll let you know." ;)
The -Xcompiler thingy is *very* necessary though, so
make try not to break it.
- Rob
--
Robert Boehne Software Engineer
Ricardo Software Chicago Technical
Bruce Korb wrote:
>
> Hi Alexandre,
How's this?
It fixes a bug, too. From the original:
> for arg
> do
> case $prev in
> "") ;;
> xcompiler)
> # Aesthetically quote the previous argument.
> prev=
> lastarg=`$echo "X$arg" | $Xsed -e "$sed_quote_subst
Hi Alexandre,
> This wouldn't work with:
>
> libtool gcc -c foo.c -Xcompiler -M
[['cuz libtool would then think ``-M'' is the source file...]]
> Automake's depcomp sometimes appends things like `-Wc,-M', but
> I'm using the equivalent `-Xcompiler -M' here because it seems
> your patch remov
>>> "adl" == Alexandre Duret-Lutz <[EMAIL PROTECTED]> writes:
[...]
adl> libtool --compile gcc -c foo.c -Xcompiler -M
Of course, I meant
libtool --mode=compile gcc -c foo.c -Xcompiler -M
[...]
--
Alexandre Duret-Lutz
___
Libtool mailing list
>>> "Bruce" == Bruce Korb <[EMAIL PROTECTED]> writes:
Bruce> As far as I can tell, the removed code serves no purpose
Bruce> (other than obfuscation ;-):
[... removal of -Xcompiler support ...]
Bruce> P.S., my head hurts
Bruce> Also, anyone care to comment about the handling of "srcfil
16 matches
Mail list logo
