| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
This has been magically working since ext re builds with -I../..,
and so picks up the inline headers from the top, the copied bogus
file has been left unused.
|
| |
|
|
| |
unpack '%65...' failures, to be more exact.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
[DELTA]
1.16: Sat 5 Apr 10:19:35 BST 2014
- Make last resort actually work from drm00
- Removed change log from Cap.pm
1.17-RC1: Wed 1 Jul 10:30:15 BST 2015
- Fix regression caused by last resort change
- Add support for capabilities of more than 2 characters
1.17: Mon 17 Aug 08:30:54 BST 2015
- No changes version bounce
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously this function unconditionally copied the OPpLVAL_INTRO and
OPpPAD_STATE flags from the LH var op to the refassign op, even when those
flag bits weren't used or meant something different.
This commit makes the copying more selective.
It also makes clear by code comments and asserts, that the refassign
op uses bit 6, OPpPAD_STATE, to mean either that or OPpOUR_INTRO
depending on the type of LHS.
I couldn't think of any test that would would break under the old regime,
but this future-proofs the code against new flags and meanings.
|
| |
|
|
|
|
|
|
| |
make the tests in the benchmark file be compilable back to 5.004_05. (To
go further back, it would need to avoid package names that start with
digits, such as 'call::sub::3_args').
Basically avoid // and our.
|
| | |
|
| |
|
|
|
| |
Since the SV is discarded almost immediately (in non-DEBUGGING builds)
don't worry about making it the smallest possible size.
|
| |
|
|
| |
This can be useful information.
|
| | |
|
| |
|
|
| |
The .md extension has become more and more standard on github
|
| | |
|
| |
|
|
|
|
|
|
| |
Follow wording from Camel book, 4th ed., p. 120, per suggestion by
Ludovic E. R. Tolhurst-Cleaver. Add Ludovic E. R. Tolhurst-Cleaver
to AUTHORS.
For: RT #125802
|
| | |
|
| |\ |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
in something like (...) = (f())[0,0]
A mortal SV can appear multiple times on the RHS without being reference
counted. If it's a string, the temptation is to steal its string buffer
during the first copy; but this would make the second usage undef.
pp_aassign() already takes account of this and uses the SV_NOSTEAL
flag when copying - except that one common code path didn't.
This commit fixes that, adds more tests, and adds more code comments
to explain the issue.
|
| | |
| |
| |
| | |
The generation number is no longer stored in the SvUVX() field
|
| | | |
|
| | |
| |
| |
| |
| | |
Since we no longer scan package vars with PL_generation for
OPpASSIGN_COMMON* purposes, eliminate the macros used for that purpose.
|
| | |
| |
| |
| |
| |
| |
| | |
Avoid setting common scalar flags in these cases:
($x) = (...);
(...) = ($x);
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
These two commits:
v5.21.3-759-gff2a62e "Skip no-common-vars optimisation for aliases"
v5.21.4-210-gc997e36 "Make list assignment respect foreach aliasing"
added a run-time mechanism to detect aliased package variables,
by either "*pkg = ...," or "for $pkg (...)", and used that information
to enable the OPpASSIGN_COMMON mechanism at runtime for detecting common
elements in a list assign, e.g.
for $alias ($a, ...) {
($a,$b) = (1,$alias);
}
The previous commit but one changed the OPpASSIGN_COMMON mechanism such
that it no longer uses PL_sawalias. So this var and the mechanism for
setting it can now be removed.
This commit removes:
* the PL_sawalias variable
* the GPf_ALIASED_SV GP flag
* the SAVEt_GP_ALIASED_SV and save_aliased_sv() save type.
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Technically in
my ($scalar,...) = @_
due to closure/goto tricks, its possible for $scalar to appear on both
the LHS and RHS, so we currently set the OPpASSIGN_COMMON_RC1 flag.
However, this imposes extra overhead; for example 5% extra instruction
reads and 11% extra conditional branches for
my ($x,$y,$z) = @_;
Given what an important construct this is, disable this flag in the
specific case of of only my's on the LHS and only @_ on the RHS.
It's technically incorrect, but its the same behaviour we've always had
(it was only the previous commit which made it safe but slower).
We still set the OPpASSIGN_COMMON_AGG flag for
my ($...,@a) = @_
since in the normal case this only adds the small additional runtime
overhead of checking that @a is already empty.
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit almost completely replaces the current mechanism
for detecting and handing common vars in list assignment, e.g.
($a,$b) = ($b,$a);
In general outline: it creates more false positives at compile-time
than before, but also no longer misses some false negatives. In
compensation, it considerably reduces the run-time cost of handling
potential and real commonality.
It does this firstly by splitting the OPpASSIGN_COMMON flag into 3
separate flags:
OPpASSIGN_COMMON_AGG
OPpASSIGN_COMMON_RC1
OPpASSIGN_COMMON_SCALAR
which indicate different classes of commonality that can be handled
in different ways at runtime.
Most importantly, it distinguishes between two basic cases. Firstly,
common scalars (OPpASSIGN_COMMON_SCALAR), e.g.
($x,....) = (....,$x,...)
where $x is modified and then sometime later its value is used again,
but that value has changed in the meantime. In this case, we need
replace such vars on the RHS with mortal copies before processing the
assign.
The second case is an aggregate on the LHS (OPpASSIGN_COMMON_AGG), e.g.
(...,@a) = (...., $a[0],...)
In this case, the issue is instead that when @a is cleared, it may free
items on the RHS (due to the stack not being ref counted). What is
required here is that rather than making of a copy of each RHS element and
storing it in the array as we progress, we make *all* the copies *before*
clearing the array, but mortalise them in case we die in the meantime.
We can further distinguish two scalar cases; sometimes it's possible
to confirm non-commonality at run-time merely by checking that all
the LHS scalars have a reference count of 1. If this is possible,
we set the OPpASSIGN_COMMON_RC1 flag rather than the
OPpASSIGN_COMMON_SCALAR flag.
The major improvement in the run-time performance in the
OPpASSIGN_COMMON_SCALAR case (or OPpASSIGN_COMMON_RC1 if rc>1 scalars are
detected), is to use a mark-and-sweep scan of the two lists using the
SVf_BREAK flag, to determine which elements are common, and only make
mortal copies of those elements. This has a very big effect on run-time
performance; for example in the classic
($a,$b) = ($b,$a);
it would formerly make temp copies of both $a and $b; now it only
copies $a.
In more detail, the mark and sweep mechanism in pp_aassign works by
looping through each LHS and RHS SV pair in parallel. It temporarily marks
each LHS SV with the SVf_BREAK flag, then makes a copy of each RHS element
only if it has the SVf_BREAK flag set. When the scan is finished, the flag
is unset on all LHS elements.
One major change in compile-time flagging is that package scalar vars are
now treated as if they could always be aliased. So we don't bother any
more to do the compile-time PL_generation checking on package vars (we
still do it on lexical vars). We also no longer make use of the run-time
PL_sawalias mechanism for detecting aliased package vars (and indeed the
next commit but one will remove that mechanism). This means that more list
assignment expressions which feature package vars will now need to
do a runtime mark-and-sweep (or where appropriate, RC1) test. In
compensation, we no longer need to test for aliasing and set PL_sawalias
in pp_gvsv and pp_gv, nor reset PL_sawalias in every pp_nextstate.
Part of the reasoning behind this is that it's nearly impossible to detect
all possible package var aliasing; for example PL_sawalias would fail to
detect XS code doing GvSV(gv) = sv.
Note that we now scan the two children of the OP_AASSIGN separately,
and in particular we mark lexicals with PL_generation only on the
LHS and test only on the RHS. So something like
($x,$y) = ($default, $default)
will no longer be regarded as having common vars.
In terms of performance, running Porting/perlbench.pl on the new
expr::aassign:: tests in t/perf/benchmarks show that the biggest slowdown
is around 13% more instruction reads and 20% more conditional branches in
this:
setup => 'my ($v1,$v2,$v3) = 1..3; ($x,$y,$z) = 1..3;',
code => '($x,$y,$z) = ($v1,$v2,$v3)',
where this is now a false positive due to the presence of package
variables.
The biggest speedup is 50% less instruction reads and conditional branches
in this:
setup => '@_ = 1..3; my ($x,$y,$z)',
code => '($x,$y,$z) = @_',
because formerly the presence of @_ pessimised things if the LHS wasn't
a my declaration (it's still pessimised, but the runtime's faster now).
Conversely, we pessimise the 'my' variant too now:
setup => '@_ = 1..3;',
code => 'my ($x,$y,$z) = @_',
this gives 5% more instruction reads and 11% more conditional branches now.
But see the next commit, which will cheat for that particular construct.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
Suggestion that we "aspire to kindness" from the ever-kind Tim Bunce.
Signed-off-by: Ricardo Signes <rjbs@cpan.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1st problem, *{"_<$filename"} globs aren't created for PP code unless
PERLDB_LINE || PERLDB_SAVESRC are on (see Perl_yylex ). Creating XSUBs
unconditionally created the *{"_<$filename"} globs for XSUB, even if PP
subs were not getting the debugging globs created. This is probably an
oversight, from commit b195d4879f which tried to deprecate using
the *{"_<$filename"} GVs for storing the originating filename of the CV
which was partially reverted 2 months later in commit 57843af05b with
CvFILE becoming a char * instead of GV *. To speed up XSUB registration
time, and decrease memory when not in Perl lang debugging mode dont create
the debugging globs for XSUBs unless in Perl lang debugging mode.
see also
http://www.nntp.perl.org/group/perl.perl5.porters/2000/06/msg13832.html
2nd problem, since the perl debugger can't step into C code, nor set
breakpoints in it, there is no reason to create *{"_<$filename"} globs
for .c files. Since someone maybe one day might try to implement that
feature, comment out the code instead of deleting it. This will slightly
reduce the size of perl binary, and speed up XSUB registration time, and
decrease memory usage when using the Perl debugger.
see also (no responses)
http://www.nntp.perl.org/group/perl.perl5.porters/2015/06/msg229014.html
perl has a number of core perma-XSUBs
(UNIVERSAL/PerlIO/DynaLoader/Internals/mro/misc const subs/etc). Each of
these previously got a "_<foo.c" glob. I counted 7 .c debugging globs on
my perl. This commit, before on WinXP, running threaded perl with
-e"system 'pause'" as a sample script, "Private Bytes" (all process
unique memory, IE not shared, not mmaped) was 488 KB, after this commit
it was 484 KB, which means that enough malloc memory was saved plus a
little bit of chance, to cross one 4 KB page of memory. IDK the exact
amount of saved memory is over or under 4KB.
|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
This commit changes to not use hard-coded code point numbers so that it
also works on EBCDIC platforms. An unintended consequence is that the
VT and \f spacing controls aren't considered a sign of binary. That
seems correct to me.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
Sleuthed by Tony Cook.
The heuristics for the 'x86' part are a bit hacky and all my fault.
|
| | |
|
| | |
|
| | |
|
| | |
|
