From: Matt S Trout <mst@shadowcat.co.uk>
Date: Mon, 21 Jun 2010 00:02:25 +0000 (+0100)
Subject: collapse threading doc changes as suggested by Nicholas, adding in notes on the addit... 
X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=commitdiff_plain;h=2d157b07717ee64a3d162a8882892d4187777823;p=p5sagit%2Fp5-mst-13.2.git

collapse threading doc changes as suggested by Nicholas, adding in notes on the additional commits related to dTHX deferral
---

diff --git a/pod/perl5132delta.pod b/pod/perl5132delta.pod
index 9d550ae..1926bf2 100644
--- a/pod/perl5132delta.pod
+++ b/pod/perl5132delta.pod
@@ -201,43 +201,23 @@ This module hasn't been updated since 1996 so we can't recommend it any more
 Only allocate entries for @_ on demand - this not only saves memory per
 subroutine defined but should hopefully improve COW behaviour (77bac2).
 
-Multiple small improvements to threads:
+=head2 Multiple small improvements to threads
 
-=over 4
-
-=item *
-
-Change the internal structured of thread->params from an SV (RV) to an AV
-- so we now pass around and store the array directly, rather than creating,
-holding and dereferencing a reference to it (78b7eff).
-
-=item *
-
-Change S_ithread_create() params from a single AV* to a pair of SV** pointers.
-This saves creating, duplicating and freeing and AV, which is only ever used for
-an internal calling convention (4cf5ea).
-
-=item *
-
-Remove redundant hv_exists() calls from ithread_create()'s spec parser (b1faab).
-
-=item *
+The internal structures of threading now make fewer API calls and fewer
+allocations, resulting in noticeably smaller object code. Additionally,
+many thread context checks have been deferred so that they're only done
+when required (although this is only possible for non-debugging builds).
 
-Skip unnecessary newAV() in ithread_create() (39f3f7).
+=head2 Eliminated xhv_fill from struct xpvhv
 
-=item *
-
-Avoid duping pads created for recursion since there's no point pre-allocating
-the same memory in the new thread (6de654).
-
-=back
+This saves 1 IV per hash and on some systems will cause struct xpvhv to
+become cache aligned. To avoid this memory saving causing a slowdown
+elsewhere, boolean use of HvFILL now calls HvTOTALKEYS instead (which
+is equivalent) - so while the fill data when actually required is now
+calculated on demand, the cases when this needs to be done should be
+few and far between (f4431c .. fcd245).
 
-Eliminated xhv_fill from struct xpvhv: This saves 1 IV per hash and on some
-systems will cause struct xpvhv to become cache aligned. To avoid this
-memory saving causing a slowdown elsewhere, boolean use of HvFILL now
-calls HvTOTALKEYS instead (which is equivalent) - so while the fill data when
-actually required is now calculated on demand, the cases when this needs to
-be done should be few and far between (f4431c .. fcd245).
+=head2 Optimisation of regexp engine string comparison work
 
 The foldEQ_utf8 API function for case-insensitive comparison of strings (which
 is used heavily by the regexp engine) was substantially refactored and