On Wed, Oct 12, 2011 at 23:36, Lawrence Crowl <crowl@google.com> wrote: > Use the mangled ...
12 years, 6 months ago
(2011-10-13 11:39:35 UTC)
#2
On Wed, Oct 12, 2011 at 23:36, Lawrence Crowl <crowl@google.com> wrote:
> Use the mangled name for merging, as this should enable us to
> handle function overloads. We use the regular identifier for other
> declarations, as that should be sufficient and avoids the problem of
> different typedefs mangling to the same name.
>
> Merge struct members as well as namespace members. This will
> eventually help with member declaration versus definition issues.
>
> Change test cases to reflect the above.
>
> Comment on other failing tests.
>
> Comment on failing cache handling for merge.
>
> Tested on x64.
>
>
> Index: gcc/testsuite/ChangeLog.pph
>
> 2011-10-12 Lawrence Crowl <crowl@google.com>
>
> * g++.dg/pph/p2pr36533.cc: Mark expected fail on unexpanded intrinsic.
> * g++.dg/pph/p4pr36533.cc: Likewise.
> * g++.dg/pph/p4mean.cc: Likewise.
> * g++.dg/pph/c3variables.cc: Comment on reason for fail.
> * g++.dg/pph/c4vardef.cc: Likewise.
>
> Index: gcc/cp/ChangeLog.pph
>
> 2011-10-12 Lawrence Crowl <crowl@google.com>
>
> * pph-streamer.h (pph_merge_name): New.
> * pph-streamer.c (pph_merge_name): New.
> * pph-streamer-out.c (pph_out_mergeable_tree_vec): Emit the vector
> in declaration order.
> (pph_out_merge_name): New.
> (pph_write_any_tree): Use pph_out_merge_name instead of raw code.
> * pph-streamer-in.c (pph_match_to_link): Use pph_merge_name.
> (pph_in_binding_level): Also merge members of structs.
> (pph_read_any_tree): Save read tree to determine if it is different
> from the tree to be used.
>
>
> Index: gcc/testsuite/g++.dg/pph/p2pr36533.cc
> ===================================================================
> --- gcc/testsuite/g++.dg/pph/p2pr36533.cc (revision 179880)
> +++ gcc/testsuite/g++.dg/pph/p2pr36533.cc (working copy)
> @@ -1,2 +1,6 @@
> /* { dg-options "-w -fpermissive" } */
> +// pph asm xdiff 25347
> +// xfail BOGUS INTRINSIC
> +// failing to recognise memset as an intrinsic
> +
> #include "p1pr36533.h"
> Index: gcc/testsuite/g++.dg/pph/c3variables.cc
> ===================================================================
> --- gcc/testsuite/g++.dg/pph/c3variables.cc (revision 179880)
> +++ gcc/testsuite/g++.dg/pph/c3variables.cc (working copy)
> @@ -1,4 +1,5 @@
> // pph asm xdiff 34997
> +// xfail BOGUS DUPVAR
> // tentative definition emitted twice
>
> #include "c0variables1.h"
> Index: gcc/testsuite/g++.dg/pph/p4mean.cc
> ===================================================================
> --- gcc/testsuite/g++.dg/pph/p4mean.cc (revision 179880)
> +++ gcc/testsuite/g++.dg/pph/p4mean.cc (working copy)
> @@ -1,4 +1,8 @@
> /* { dg-options "-w -fpermissive" } */
> +// pph asm xdiff 39234
> +// xfail BOGUS INTRINSIC
> +// failing to recognize sqrt as an intrinsic
> +
> #include <stdlib.h>
> #include <stdio.h>
> #include <math.h>
> Index: gcc/testsuite/g++.dg/pph/p4pr36533.cc
> ===================================================================
> --- gcc/testsuite/g++.dg/pph/p4pr36533.cc (revision 179880)
> +++ gcc/testsuite/g++.dg/pph/p4pr36533.cc (working copy)
> @@ -1,2 +1,6 @@
> /* { dg-options "-w -fpermissive" } */
> +// pph asm xdiff 25347
> +// xfail BOGUS INTRINSIC
> +// failing to recognise memset as an intrinsic
> +
> #include "p4pr36533.h"
> Index: gcc/testsuite/g++.dg/pph/c4vardef.cc
> ===================================================================
> --- gcc/testsuite/g++.dg/pph/c4vardef.cc (revision 179880)
> +++ gcc/testsuite/g++.dg/pph/c4vardef.cc (working copy)
> @@ -1,4 +1,6 @@
> // pph asm xdiff 00553
> +// xfail BOGUS DUPVAR
> +// definition emitted twice
>
> #include "c0vardef1.h"
> #include "c0vardef2.h"
> Index: gcc/cp/pph-streamer-in.c
> ===================================================================
> --- gcc/cp/pph-streamer-in.c (revision 179880)
> +++ gcc/cp/pph-streamer-in.c (working copy)
> @@ -803,7 +803,7 @@ pph_match_to_function (tree expr ATTRIBU
> against an LINK of a chain. */
>
> static tree
> -pph_match_to_link (tree expr, location_t where, const char *idstr, tree*
link)
> +pph_match_to_link (tree expr, location_t where, const char *idstr, tree
*link)
> {
> enum tree_code link_code, expr_code;
> tree idtree;
> @@ -817,7 +817,7 @@ pph_match_to_link (tree expr, location_t
> if (link_code != expr_code)
> return NULL;
>
> - idtree = DECL_NAME (*link);
> + idtree = pph_merge_name (*link);
> if (!idtree)
> return NULL;
>
> @@ -1072,7 +1072,7 @@ pph_in_binding_level (cp_binding_level *
> *out_field = bl;
>
> entity = bl->this_entity = pph_in_tree (stream);
> - if (NAMESPACE_SCOPE_P (entity))
> + if (NAMESPACE_SCOPE_P (entity) || DECL_CLASS_SCOPE_P (entity))
> {
> if (flag_pph_debug >= 3)
> debug_tree_chain (bl->names);
> @@ -1962,7 +1962,8 @@ pph_read_any_tree (pph_stream *stream, t
> {
> struct lto_input_block *ib = stream->encoder.r.ib;
> struct data_in *data_in = stream->encoder.r.data_in;
> - tree expr = NULL_TREE;
> + tree read = NULL;
> + tree expr = NULL;
> enum pph_record_marker marker;
> unsigned image_ix, ix;
> enum LTO_tags tag;
> @@ -1998,9 +1999,11 @@ pph_read_any_tree (pph_stream *stream, t
>
> /* Materialize a new node from IB. This will also read all the
> language-independent bitfields for the new tree. */
> - expr = pph_read_tree_header (stream, tag);
> + expr = read = pph_read_tree_header (stream, tag);
> + gcc_assert (read != NULL);
> if (chain)
> expr = pph_merge_into_chain (stream, expr, chain);
> + gcc_assert (expr != NULL);
> }
>
> gcc_assert (marker == PPH_RECORD_START
> @@ -2021,6 +2024,7 @@ pph_read_any_tree (pph_stream *stream, t
> /* Add the new tree to the cache and read its body. The tree
> is added to the cache before we read its body to handle
> circular references and references from children nodes. */
> + /* FIXME pph: We should not insert when read == expr, but it fails. */
No. We should always insert this tree in the cache. Otherwise,
subsequent IREF records will not find what they're looking for. If
READ == EXPR, it means that we did not do the merge, right?
pph_read_tree_header will allocate an incomplete tree and assign it to
EXPR and READ, so the only way we should be getting the same tree back
from pph_merge_into_chain is if it gives us back the same tree we just
allocated.
Regardless of what we get here, the tree we materialize must be added
to the cache.
> + tree name = pph_merge_name (expr);
> + if (name)
> + pph_out_string_with_length (stream, IDENTIFIER_POINTER (name),
> + IDENTIFIER_LENGTH (name));
Indent here needs to align with the open braces above.
I'm seeing an infinite loop in g++.dg/pph/c1limits-externalid.cc. The while() loop in pph_search_in_chain is not ending. ...
12 years, 6 months ago
(2011-10-13 15:18:10 UTC)
#3
I'm seeing an infinite loop in g++.dg/pph/c1limits-externalid.cc. The
while() loop in pph_search_in_chain is not ending. Or maybe it's
falling into the N^2 trap you mention in that routine?
I've added a short timeout to this test and XFAIL'd it so you can debug it.
Diego.
Issue 5268042: [pph] More DECL merging.
(Closed)
Created 12 years, 6 months ago by Lawrence Crowl
Modified 11 years, 11 months ago
Reviewers: Diego Novillo
Base URL: svn+ssh://gcc.gnu.org/svn/gcc/branches/pph/
Comments: 0
