* [C++ PATCH] Handle TYPE_PACK_EXPANSION return in tsubst_exception_specification
@ 2007-11-18 15:46 Doug Gregor
2007-11-18 21:51 ` Jakub Jelinek
2007-12-03 21:39 ` Jason Merrill
0 siblings, 2 replies; 4+ messages in thread
From: Doug Gregor @ 2007-11-18 15:46 UTC (permalink / raw)
To: GCC Patches
[-- Attachment #1: Type: text/plain, Size: 670 bytes --]
This little patch fixes PR 33509, a P2 ice-on-valid where we failed to
handle the instantiation of exception specifications that involve
parameter packs from multiple levels of template parameters. The fix
is relatively simple; we just deal with the resulting
TYPE_PACK_EXPANSION explicitly.
Tested i686-pc-linux-gnu; okay for mainline?
- Doug
2007-11-18 Douglas Gregor <doug.gregor@gmail.com>
PR c++/33509
* pt.c (tsubst_exception_specification): Handle substitutions into
member templates, where tsubst_pack_expansion returns a
TYPE_PACK_EXPANSION.
2007-11-18 Douglas Gregor <doug.gregor@gmail.com>
PR c++/33509
* g++.dg/cpp0x/variadic-throw.C: New.
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: variadic-throw.patch --]
[-- Type: text/x-patch; name=variadic-throw.patch, Size: 1662 bytes --]
Index: cp/pt.c
===================================================================
--- cp/pt.c (revision 130268)
+++ cp/pt.c (working copy)
@@ -8653,7 +8653,24 @@ tsubst_exception_specification (tree fnt
expanded_specs = tsubst_pack_expansion (TREE_VALUE (specs),
args, complain,
in_decl);
- len = TREE_VEC_LENGTH (expanded_specs);
+
+ if (expanded_specs == error_mark_node)
+ return error_mark_node;
+ else if (TREE_CODE (expanded_specs) == TREE_VEC)
+ len = TREE_VEC_LENGTH (expanded_specs);
+ else
+ {
+ /* We're substituting into a member template, so
+ we got a TYPE_PACK_EXPANSION back. Add that
+ expansion and move on. */
+ gcc_assert (TREE_CODE (expanded_specs)
+ == TYPE_PACK_EXPANSION);
+ new_specs = add_exception_specifier (new_specs,
+ expanded_specs,
+ complain);
+ specs = TREE_CHAIN (specs);
+ continue;
+ }
}
for (i = 0; i < len; ++i)
Index: testsuite/g++.dg/cpp0x/variadic-throw.C
===================================================================
--- testsuite/g++.dg/cpp0x/variadic-throw.C (revision 0)
+++ testsuite/g++.dg/cpp0x/variadic-throw.C (revision 0)
@@ -0,0 +1,20 @@
+// { dg-options -std=c++0x }
+// PR c++/33509
+template<int M, int N> struct pair
+{
+ int i, j;
+ pair() : i(M), j(N) {}
+};
+
+template<int... M> struct S
+{
+ template<int... N> static int foo() throw (pair <M, N>...)
+ {
+ return 1;
+ }
+};
+
+int bar ()
+{
+ return S<0, 1, 2>::foo<0, 1, 3> ();
+}
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [C++ PATCH] Handle TYPE_PACK_EXPANSION return in tsubst_exception_specification
2007-11-18 15:46 [C++ PATCH] Handle TYPE_PACK_EXPANSION return in tsubst_exception_specification Doug Gregor
@ 2007-11-18 21:51 ` Jakub Jelinek
2007-11-19 16:43 ` Doug Gregor
2007-12-03 21:39 ` Jason Merrill
1 sibling, 1 reply; 4+ messages in thread
From: Jakub Jelinek @ 2007-11-18 21:51 UTC (permalink / raw)
To: Doug Gregor; +Cc: GCC Patches
On Sun, Nov 18, 2007 at 09:38:27AM -0500, Doug Gregor wrote:
> This little patch fixes PR 33509, a P2 ice-on-valid where we failed to
> handle the instantiation of exception specifications that involve
> parameter packs from multiple levels of template parameters. The fix
> is relatively simple; we just deal with the resulting
> TYPE_PACK_EXPANSION explicitly.
>
> Tested i686-pc-linux-gnu; okay for mainline?
>
> 2007-11-18 Douglas Gregor <doug.gregor@gmail.com>
>
> PR c++/33509
> * pt.c (tsubst_exception_specification): Handle substitutions into
> member templates, where tsubst_pack_expansion returns a
> TYPE_PACK_EXPANSION.
>
> 2007-11-18 Douglas Gregor <doug.gregor@gmail.com>
>
> PR c++/33509
> * g++.dg/cpp0x/variadic-throw.C: New.
I think it would be good to add not only this valid testcase, but
the invalid one as well (i.e. one with
return S<0, 1, 2>::foo<0, 1> ();
rather than
return S<0, 1, 2>::foo<0, 1, 3> ();
) to make sure it doesn't ICE and issues some meaningful error
on invalid code.
Jakub
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [C++ PATCH] Handle TYPE_PACK_EXPANSION return in tsubst_exception_specification
2007-11-18 21:51 ` Jakub Jelinek
@ 2007-11-19 16:43 ` Doug Gregor
0 siblings, 0 replies; 4+ messages in thread
From: Doug Gregor @ 2007-11-19 16:43 UTC (permalink / raw)
To: Jakub Jelinek; +Cc: GCC Patches
On Nov 18, 2007 3:01 PM, Jakub Jelinek <jakub@redhat.com> wrote:
> > 2007-11-18 Douglas Gregor <doug.gregor@gmail.com>
> >
> > PR c++/33509
> > * g++.dg/cpp0x/variadic-throw.C: New.
>
> I think it would be good to add not only this valid testcase, but
> the invalid one as well (i.e. one with
> return S<0, 1, 2>::foo<0, 1> ();
> rather than
> return S<0, 1, 2>::foo<0, 1, 3> ();
> ) to make sure it doesn't ICE and issues some meaningful error
> on invalid code.
Okay, will do.
- Doug
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [C++ PATCH] Handle TYPE_PACK_EXPANSION return in tsubst_exception_specification
2007-11-18 15:46 [C++ PATCH] Handle TYPE_PACK_EXPANSION return in tsubst_exception_specification Doug Gregor
2007-11-18 21:51 ` Jakub Jelinek
@ 2007-12-03 21:39 ` Jason Merrill
1 sibling, 0 replies; 4+ messages in thread
From: Jason Merrill @ 2007-12-03 21:39 UTC (permalink / raw)
To: Doug Gregor; +Cc: GCC Patches
OK.
Jason
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2007-12-03 21:39 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-11-18 15:46 [C++ PATCH] Handle TYPE_PACK_EXPANSION return in tsubst_exception_specification Doug Gregor
2007-11-18 21:51 ` Jakub Jelinek
2007-11-19 16:43 ` Doug Gregor
2007-12-03 21:39 ` Jason Merrill
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).