On Mon, Feb 26, 2001 at 08:04:29PM -0300, Alexandre Oliva wrote:
> On Feb 24, 2001, Sam TH <[EMAIL PROTECTED]> wrote:
>
> > On Sun, Feb 25, 2001 at 12:18:13AM +0100, Peter Eisentraut wrote:
>
> >> It seems this was fully intentional. In fact, the code says
> >>
> >> # Make sure that $var contains only unique libraries
> >> # and add them in reverse order
>
> This is definitely wrong. Thanks for spotting the problem.
>
> > Is there any good reason to prevent people from doing this?
>
> The rationale was that it used to prevent correct linking because of
> excessively-long link command lines in certain cases. Which doesn't
> make the dropping of duplicate libraries right, it just creates a
> different problem scenario.
>
> Sam, would you please add a `break' just before the comment above, and
> see if it fixes the problem for you?
>
I did this. It does fix that problem.
However, it has the following 2 problems (right offhand)
1. It causes -lfoo portions of the line that are added by libtool to
be duplicated. So my previously posted example ends up with
-lm ./.libs/libfoo.a -lm ./.libs/libfoo.a
2. It causes the following 8 tests to fail. (I don't think they
failed before.)
FAIL: depdemo-make.test
FAIL: depdemo-make.test
FAIL: depdemo-exec.test
FAIL: depdemo-inst.test
FAIL: link.test
FAIL: depdemo-make.test
FAIL: depdemo-exec.test
FAIL: depdemo-inst.test
That's in order.
sam th
[EMAIL PROTECTED]
http://www.abisource.com/~sam/
GnuPG Key:
http://www.abisource.com/~sam/key
PGP signature