From: "Mehmet Yavuz Selim Soyturk" (via RT) <[EMAIL PROTECTED]>
Date: Wed, 28 Feb 2007 08:54:19 -0800
# New Ticket Created by "Mehmet Yavuz Selim Soyturk"
# Please include the string: [perl #41642]
# in the subject line of all future correspondence about this issue.
# <URL: http://rt.perl.org/rt3/Ticket/Display.html?id=41642 >
That program causes me segfault. I only did get a segfault when:
1) there is a tailcall _method_
2) the method gets its parameters with :slurpy
I don't get any error if I invoke parrot with the -G flag . . .
Does the following patch fix it? If so, and this patch does work for me
in r17222, then this is the same problem I discovered on 25-Dec-06:
From: Bob Rogers <[EMAIL PROTECTED]>
Subject: GC problem in parrot_pass_args to a tailcall (r16239)
Date: Mon, 25 Dec 2006 15:30:32 -0500
There is a repeatable segfault when GC is triggered during argument
passing after a tailcall. The immediate symptom is that clone_key_arg
blows cookies because the current arg is a 0xdeadbeef PMC. Apparently,
the sweep doesn't see pointers in the old context . . .
My test case didn't require a method, but I also couldn't reproduce it
reliably without a code tweak to force a GC during arg processing.
I still don't know what the "real" fix ought to look like. It wasn't
obvious to me at the time . . . and then I got too busy.
-- Bob Rogers
http://rgrjr.dyndns.org/
Diffs between last version checked in and current workfile(s):
Index: src/ops/core.ops
===================================================================
--- src/ops/core.ops (revision 17060)
+++ src/ops/core.ops (working copy)
@@ -538,7 +538,9 @@
interp->current_args = NULL;
interp->current_params = NULL;
+ Parrot_block_DOD(interp);
parrot_pass_args(interp, caller_ctx, ctx, src_indexes, dst_indexes,
PARROT_PASS_PARAMS);
+ Parrot_unblock_DOD(interp);
if (PObj_get_FLAGS(ccont) & SUB_FLAG_TAILCALL) {
PObj_get_FLAGS(ccont) &= ~SUB_FLAG_TAILCALL;
--ctx->recursion_depth;
End of diffs.
