Commit e8ba94fc by Jan Hubicka Committed by Jan Hubicka

re PR c/21659 ([unit-at-a-time] "weak declaration must precede definition" error missing at >= O1)


	PR c/21659
	* doc/extend.texi (weak pragma): Drop claim that it must
	appear before definition.
	* varasm.c (merge_weak, declare_weak): Only sanity check
	that DECL is not output at a time it is declared weak.

From-SVN: r169184
parent 71c67403
2011-01-24 Jan Hubicka <jh@suse.cz>
PR c/21659
* doc/extend.texi (weak pragma): Drop claim that it must
appear before definition.
* varasm.c (merge_weak, declare_weak): Only sanity check
that DECL is not output at a time it is declared weak.
2011-01-24 Kenneth Zadeck <zadeck@naturalbridge.com>
* machmode.def: Fixed comments.
......
......@@ -13024,8 +13024,7 @@ aliases.
@cindex pragma, weak
This pragma declares @var{symbol} to be weak, as if the declaration
had the attribute of the same name. The pragma may appear before
or after the declaration of @var{symbol}, but must appear before
either its first use or its definition. It is not an error for
or after the declaration of @var{symbol}. It is not an error for
@var{symbol} to never be defined at all.
@item #pragma weak @var{symbol1} = @var{symbol2}
......
......@@ -5139,20 +5139,16 @@ merge_weak (tree newdecl, tree olddecl)
/* NEWDECL is weak, but OLDDECL is not. */
/* If we already output the OLDDECL, we're in trouble; we can't
go back and make it weak. This error cannot be caught in
declare_weak because the NEWDECL and OLDDECL was not yet
been merged; therefore, TREE_ASM_WRITTEN was not set. */
if (TREE_ASM_WRITTEN (olddecl))
error ("weak declaration of %q+D must precede definition",
newdecl);
go back and make it weak. This should never happen in
unit-at-a-time compilation. */
gcc_assert (!TREE_ASM_WRITTEN (olddecl));
/* If we've already generated rtl referencing OLDDECL, we may
have done so in a way that will not function properly with
a weak symbol. */
else if (TREE_USED (olddecl)
&& TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (olddecl)))
warning (0, "weak declaration of %q+D after first use results "
"in unspecified behavior", newdecl);
a weak symbol. Again in unit-at-a-time this should be
impossible. */
gcc_assert (!TREE_USED (olddecl)
|| !TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (olddecl)));
if (TARGET_SUPPORTS_WEAK)
{
......@@ -5184,10 +5180,9 @@ merge_weak (tree newdecl, tree olddecl)
void
declare_weak (tree decl)
{
gcc_assert (TREE_CODE (decl) != FUNCTION_DECL || !TREE_ASM_WRITTEN (decl));
if (! TREE_PUBLIC (decl))
error ("weak declaration of %q+D must be public", decl);
else if (TREE_CODE (decl) == FUNCTION_DECL && TREE_ASM_WRITTEN (decl))
error ("weak declaration of %q+D must precede definition", decl);
else if (!TARGET_SUPPORTS_WEAK)
warning (0, "weak declaration of %q+D not supported", decl);
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment