This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix PR 18746, fab causing more problems

On Dec 1, 2004, at 5:19 PM, Andrew MacLeod wrote:

On Wed, 2004-12-01 at 17:12, Diego Novillo wrote:
On Wed, 2004-12-01 at 16:54 -0500, Andrew Pinski wrote:

	* tree-ssa-alias.c (init_alias_info): Get statement operands for all
	the statements if aliases_computed_p is true.

The comment I had suggested on IRC is more descriptive. Don't recall it
exactly. If you still have it in your scrollback buffer, please use
that. Otherwise, it was something like:

<dnovillo> {
<dnovillo> unsigned i;
<dnovillo> bitmap_iterator bi;
<dnovillo> + basic_block bb;
<dnovillo> + block_stmt_iterator si;
<dnovillo> +
<dnovillo> + /* Make sure that every statement has a valid set of operands.
<dnovillo> + If a statement needs to be scanned for operands while we
<dnovillo> + compute aliases, it may get erroneous operands because all
<dnovillo> + the alias relations are not built at that point. */
<dnovillo> + FOR_EACH_BB (bb)
<dnovillo> + for (si = bsi_start (bb); !bsi_end_p (si); bsi_next (&si))
<dnovillo> + get_stmt_operands (bsi_stmt (si));

I also added the following comment to the above one as we should not need
this when Andrew MacLeod's work for updating operands when modify_stmt is
FIXME: This code will become obsolete when operands are not
lazily updated.

Andrew Pinski

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]