Much fallout from PR 27668

Mark Mitchell mark@codesourcery.com
Tue Aug 8 08:14:00 GMT 2006


Lee Millward wrote:

> When Mark approved the patch he predicted there would be some fallout,
> I've been keeping a lookout for any new regressions which have
> resulted from my patch and trying to fix them as I come across them.

I think that Volker's right.  Let's back the original patch out of 4.1,
and leave well enough alone; let's just pursue this on the mainline.

>        PR c++/28637
>        * pt.c (coerce_template_parms): Copy across the
>        invalid template arguments to the new template inner arguments.

OK.

>        (retrieve_specialization): Robustify.

OK.

>        PR c++/28638
>        * pt.c (coerce_template_template_parms): Robustify.

OK.

>        PR c++/28639
>        * error.c (dump_template_parms): Robustify.

We should not silently skip the error_mark_node here.  Instead, we
should print "<template parameter error>" as in the NULL argument case.
 (That should never occur; an erroneous argument should be
error_mark_node, but that's another issue...)  OK with that change.

>        PR c++/28640
>        * pt.c (redeclare_class_template): Robustify.

OK.

>        PR c++/28641
>        * pt.c (type_unification_real): Robustify.

OK.

-- 
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713



More information about the Gcc-patches mailing list