* [C++ Patch] PR 52987
@ 2015-07-22 22:19 Paolo Carlini
2015-07-23 7:46 ` Jason Merrill
0 siblings, 1 reply; 2+ messages in thread
From: Paolo Carlini @ 2015-07-22 22:19 UTC (permalink / raw)
To: gcc-patches; +Cc: Jason Merrill
[-- Attachment #1: Type: text/plain, Size: 982 bytes --]
Hi,
this bug is purely about error recovery. A while ago I fixed the first
half, but for, eg:
int foo(x a) {
}
we still emit the pointless:
52987_2.C:1:14: error: expected â,â or â;â before â{â token
In fact, we *already* have code helping error recovery in
cp_parser_simple_declaration:
/* If we have already issued an error message we don't need
to issue another one. */
if (decl != error_mark_node
|| cp_parser_uncommitted_to_tentative_parse_p (parser))
cp_parser_error (parser, "expected %<,%> or %<;%>");
but it doesn't trigger in such cases, because the decl is in fact !=
error_mark_node, trace of the error can be found only in its
DECL_INITIAL (I noticed that only today ;) Thus the below, which so far
appears to work well for me, passes testing on x86_64-linux
(g++.old-deja/g++.law/init7.C included, which blocks even simpler
solutions).
Thanks,
Paolo.
////////////////////////
[-- Attachment #2: CL_52987 --]
[-- Type: text/plain, Size: 318 bytes --]
/cp
2015-07-23 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/52987
* parser.c (cp_parser_simple_declaration): Robustify check avoiding
duplicated error messages.
/testsuite
2015-07-23 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/52987
* g++.dg/parse/error57.C: New.
* g++.dg/expr/string-2.C: Update.
[-- Attachment #3: patch_52987 --]
[-- Type: text/plain, Size: 1298 bytes --]
Index: cp/parser.c
===================================================================
--- cp/parser.c (revision 226075)
+++ cp/parser.c (working copy)
@@ -11660,7 +11660,8 @@ cp_parser_simple_declaration (cp_parser* parser,
{
/* If we have already issued an error message we don't need
to issue another one. */
- if (decl != error_mark_node
+ if ((decl != error_mark_node
+ && DECL_INITIAL (decl) != error_mark_node)
|| cp_parser_uncommitted_to_tentative_parse_p (parser))
cp_parser_error (parser, "expected %<,%> or %<;%>");
/* Skip tokens until we reach the end of the statement. */
Index: testsuite/g++.dg/expr/string-2.C
===================================================================
--- testsuite/g++.dg/expr/string-2.C (revision 226075)
+++ testsuite/g++.dg/expr/string-2.C (working copy)
@@ -4,7 +4,7 @@
char a[1];
int foo(a = "") // { dg-error "invalid array assignment" }
-{ // { dg-error "" }
+{
return 0;
}
Index: testsuite/g++.dg/parse/error57.C
===================================================================
--- testsuite/g++.dg/parse/error57.C (revision 0)
+++ testsuite/g++.dg/parse/error57.C (working copy)
@@ -0,0 +1,4 @@
+// PR c++/52987
+
+int foo(x a) { // { dg-error "9:'x' was not declared in this scope" }
+}
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [C++ Patch] PR 52987
2015-07-22 22:19 [C++ Patch] PR 52987 Paolo Carlini
@ 2015-07-23 7:46 ` Jason Merrill
0 siblings, 0 replies; 2+ messages in thread
From: Jason Merrill @ 2015-07-23 7:46 UTC (permalink / raw)
To: Paolo Carlini, gcc-patches
OK.
Jason
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2015-07-23 3:59 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-07-22 22:19 [C++ Patch] PR 52987 Paolo Carlini
2015-07-23 7:46 ` 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).