summaryrefslogtreecommitdiff
path: root/cop.h
Commit message (Collapse)AuthorAgeFilesLines
* Groundwork to allow cops and pmops to store the UTF8 flagBrian Fraser2011-10-061-3/+15
| | | | | | | | | With threaded builds, cop.h and op.h get an extra member in their structs, to save the UTF-8ness of the stash's name. *STASH_set() checks for the flag, stores it through *STASH_flags(), and *STASH() uses the latter to fetch the correct scalar.
* remove index offsetting ($[)Zefram2011-09-091-24/+0
| | | | | | $[ remains as a variable. It no longer has compile-time magic. At runtime, it always reads as zero, accepts a write of zero, but dies on writing any other value.
* Use OPpDEREF for lvalue sub, such that the flags contains the deref type, ↵Gerard Goossen2011-09-011-1/+1
| | | | | | | instead of deriving it from the opchain. Also contains a test where using the opchain to determine the deref type fails.
* Rename store/fetch_cop_label as cop_*Father Chrysostomos2011-07-161-1/+1
| | | | | | | | This makes them consistent with other functions that put the basic datum type first (like hv_*, sv_*, cophh_*). Since fetch_cop_label is marked as experimental (M), this change should be OK.
* Change was_lvalue_sub back to X; spell out the only use of itFather Chrysostomos2011-07-141-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | OK, now I understand what’s happening. If there is a public macro (PUSHSUB) that contains a call to a pri- vate function (was_lvalue_sub), that function has to be exported, so that non-core code can call it. But if it is marked X, there is no was_lvalue_sub shorthand macro visible to non-core code, so when the PUSHSUB macro is expanded in such code, the was_lvalue_sub(...) bit becomes a call to the function literally named was_lvalue_sub, as opposed to Perl_lvalue_sub (and is compiled that way on forgiving platforms). Making it A makes that macro available to non-core code, but also implies that it is available for direct use by extensions, which is not the case with was_lvalue_sub. So, this commit makes it X again, but spells it out in PUSHSUB, so there is no need for the function’s macro to be available when PUSHSUB is expanded. Hence, there is no need for the was_lvalue_sub macro to exist, so this commit also removes it. See also these three commits: c73b0699db 7b70e81778 777d901444
* Propagate (non-)lvalue context through nested callsFather Chrysostomos2011-07-091-1/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Before this commit, this code would fail: $foo = "foo"; sub foo :lvalue{ return index "foo","o" } sub bar :lvalue { foo } $x = bar; (It would fail for ‘return $]’ as well. Whether it’s a PADTMP or a read-only scalar makes no difference.) foo would think it was being called in true lvalue context, because the entersub op that called it (in bar) was marked that way, bar being an lvalue sub as well. The PUSHSUB macro in cop.h needed to be modified to account for dynamic, or indetermine, context (i.e., indeterminable at compile time). This happens when an entersub op is an argument to return or the last statement in a subroutine. In those cases it has to propa- gate the context from the caller. So what we now do is this: Both lvalue and in-args flags are turned on for an entersub op when op_lvalue is called with OP_LEAVESUBLV as the type. Then PUSHSUB copies into the context stack only those flags that are set both on the current entersub op and in the context stack for the previous sub call.
* [perl #7946] Lvalue subs do not autovivifyFather Chrysostomos2011-06-031-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This commit makes autovivification work with lvalue subs. It follows the same technique used by other autovivifiable ops (aelem, helem, tc.), except that, due to flag constraints, it uses a single flag and instead checks the op tree at run time to find out what sort of thing to vivify. The flag constraints are that these two flags: #define OPpENTERSUB_HASTARG 32 /* Called from OP tree. */ #define OPpENTERSUB_NOMOD 64 /* Immune to op_lvalue() for :attrlist. */ conflict with these: #define OPpDEREF (32|64) /* autovivify: Want ref to something: */ #define OPpDEREF_AV 32 /* Want ref to AV. */ #define OPpDEREF_HV 64 /* Want ref to HV. */ #define OPpDEREF_SV (32|64) /* Want ref to SV. */ Renumbering HASTARG and NOMOD is problematic, as there are places in op.c that change entersubs into rv2cvs, and the entersub and rv2cv flags would conflict. Setting the flags correctly when changing the type is hard and would result in subtle bugs if not done perfectly. Ops like ${...} don’t actually autovivify; it’s the op inside that does it. In those cases, the parent op is flagged with OPpDEREFed, and it skips get-magic, as it has already been called by the inner op. Since entersub is now marked as being an autovivifying op, ${...} in lvalue context ends up skipping get-magic if there is a foo() inside. And this affects even regular subs. So pp_leavesub and pp_return have to call get-magic; hence the new tests in gmagic.t.
* cop.h: pod: Fix broken linksKarl Williamson2011-05-181-4/+4
|
* struct subst; remove macro for obsolete fieldDavid Mitchell2011-02-181-1/+0
| | | | The field was removed a while ago, but the macro sb_once remained.
* Update old activestate links to point to git.Michael Stevens2011-01-201-2/+3
|
* In Perl_write_to_stderr(), use Perl_magic_methcall() if STDERR is tied.Nicholas Clark2011-01-131-0/+2
| | | | | | Add a flag G_WRITING_TO_STDERR to signal that Perl_magic_methcall() needs to localise PL_stderrgv to NULL, and save/free temps, inside its ENTER/LEAVE pair.
* Fix typos (spelling errors) in Perl sources.Peter J. Acklam) (via RT2011-01-071-1/+1
| | | | | | | | | # New Ticket Created by (Peter J. Acklam) # Please include the string: [perl #81904] # in the subject line of all future correspondence about this issue. # <URL: http://rt.perl.org/rt3/Ticket/Display.html?id=81904 > Signed-off-by: Abigail <abigail@abigail.be>
* [perl #80548] Add the stash name to DTrace probesDavid Leadbeater2010-12-101-2/+4
| | | | | | | | | | | This adds an additional parameter to perl's dtrace probes with the stash name of the subroutine. This generally looks nicer than the filename but gives a similar level of context. As this is an additional parameter this will not have an impact on existing DTrace scripts. (Also due to the way DTrace works I believe it does not break binary compatibility and would be safe to backport to maint-5.12 if desired, but I'm not a DTrace expert.)
* mark cophh API as experimentalZefram2010-10-211-16/+16
|
* full API for cop hint hashesZefram2010-10-211-7/+303
| | | | | | | | | | | | | Expose cop hint hashes as a type COPHH, with a cophh_* API which is a macro layer over the refcounted_he_* API. The documentation for cophh_* describes purely API-visible behaviour, whereas the refcounted_he_* documentation describes the functions mainly in terms of the implementation. Revise the cop_hints_* API, using the flags parameter consistently and reimplementing in terms of cophh_*. Use the cophh_* and cop_hints_* functions consistently where appropriate. [Modified by the committer to update two calls to Perl_refcounted_he_fetch recently added to newPMOP.]
* Recursive MULTICALL prematurely freed CVDavid Mitchell2010-10-191-2/+2
| | | | | | | | See [perl #78070]. Basically, POPSUB/LEAVESUB had a mechanism to decrement the reference count of the CV only at CvDEPTH==1; POP_MULTICALL was decrementing it at all depths.
* Define CxPADLOOP unconditionally, as post f83b46a0 it is always used.Nicholas Clark2010-09-091-3/+1
| | | | Previously it was only used under -DITHREADS
* bad things happened with for $x (...) { *x = *y }David Mitchell2010-09-081-5/+9
| | | | | | | | | | | | | | | | | | | | | | | | fix for [perl #21469]: since the GP may be pulled from under us and freed, coredumps and strange things can happen. Fix this by storing a pointer to the GV in the loop block, rather than a pointer to the GvSV slot. The ITHREADS variant already stores GV rather than than &GvSV; extend this to non-threaded builds too. Also, for both threaded and non-threaded, it used to push &GvSV on the save stack. Fix this by introducing a new save type, SAVEt_GVSV. This behaves similarly to SAVEt_SV, but without magic get/set. This means that for $package_var (...) is now close in behaviour to local $package_var = ... (except for the magic bit).
* create itervar_u union in struct block_loopDavid Mitchell2010-09-081-19/+12
| | | | | | | make it clearer what types of pointer to the iterator variable can be stored, reduce the amount of #ifdef USE_ITHREADS, get rid of some macros, and generally make the code easier to follow. No change to the size of the structure.
* eliminate targoffset from struct block_loopDavid Mitchell2010-09-081-6/+4
| | | | This value is also available via via cx->blk_loop.my_op->op_targ
* eliminate next_op from struct block_loopDavid Mitchell2010-09-081-13/+0
| | | | | | This field is only used in non-threaded builds, and the comments imply that this is because in non-threaded builds this value may be modified. But nothing in core modifies it.
* Change the first argument of Perl_fetch_cop_label() to COP *Nicholas Clark2010-09-021-1/+1
| | | | | | | | | | From a suggestion from Ben Morrow. The first argument used to be struct refcounted_he *, which exposed an implementation detail - that the COP's labels are (now) stored in this way. Google Code Search and an unpacked CPAN both fail to find any users of this API, so the impact should be minimal.
* Revert "Fix off-by-one: avoid allocating an extra context"Chip Salzenberg2010-07-271-1/+1
| | | | | This reverts commit 395b8e2d02eadc9b0639534410c39c530bc8a33d. The fencepost error is coming from inside the programmer!
* Fix off-by-one: avoid allocating an extra contextChip Salzenberg2010-07-271-1/+1
| | | | (patch req by Nicholas)
* Stop using WITH_THR and WITH_THX, as they were never necessary here.Nicholas Clark2010-06-161-2/+2
|
* Merge remote branch 'zefram/zefram/reliable_exception' into bleadRafael Garcia-Suarez2010-05-041-1/+1
|\ | | | | | | | | Conflicts: pp_ctl.c
| * bring G_KEEPERR back to the realm of sanityZefram2010-04-231-1/+1
| | | | | | | | | | | | | | | | | | Makes the G_KEEPERR logic more consistent, and in particular make it sensibly handle non-string exceptions. An exception in a destructor is now always emitted as a warning, and never copied or merged into $@ of the surrounding context. No more clobbering exceptions being handled elsewhere, and no more double reporting. This fixes the rest of [perl #74538].
* | For Perl_magic_methcall() add G_UNDEF_FILL to fill the stack with &PL_sv_undef.Nicholas Clark2010-04-261-0/+3
|/ | | | | | This replaces the previous special case of using a negative argument count to signify this, allowing the argument count to become unsigned. Rename it from n to argc.
* Move PERL_ASYNC_CHECK() from POPBLOCK() to the kill case of Perl_apply().Nicholas Clark2010-04-181-1/+0
| | | | | | This ensures that (safe) signals sent to the same process are still dispatched within the same statement (as before), without overloading the semantics of block popping.
* Move PERL_ASYNC_CHECK() from the runloop to control flow OPs.Nicholas Clark2010-04-151-0/+1
| | | | | | | | | For the typical code this gives a 5% speedup, and removes the cost of "safe signals". Tight looping code will show less gains, but should never be slower. Subtle bugs might remain - there might be constructions that enter the runloop (where signals used to be dispatched) but don't contain any PERL_ASYNC_CHECK() calls themselves.
* more mods to -Dl debugging outputDavid Mitchell2010-04-061-9/+6
|
* improve -Dl debugging outputDavid Mitchell2010-03-301-21/+61
| | | | | In particular, distinguish between scope and context stack push/pops, show depth of JUMPENV stack, and show STACKINFO push/pops
* Eliminate OP_SETSTATE from cop.h headerReini Urban2009-12-131-3/+3
| | | | | | | | | It had been added with change 3728 to track linenumbers in optimized else, disabled by change 4309, and removed with change 33072. Bump copyright, latest change was "Fix MULTICALL in List-Util" 2009-03-07 with commit 1bbbfc50
* Add line information to jumplevel debug informationGerard Goossen2009-11-121-4/+6
| | | | | | | | | | | | | Add information about where in the C code the jumplevel poping/setting up was done. Gerard From 7b95a19d6fbd3615a034cea79fa087b80e4a9555 Mon Sep 17 00:00:00 2001 From: Gerard Goossen <gerard@ggoossen.net> Date: Thu, 12 Nov 2009 16:50:13 +0100 Subject: [PATCH] Add line information to jumplevel debug information provided when using -Dl Signed-off-by: H.Merijn Brand <h.m.brand@xs4all.nl>
* Add assertion to JMPENV_POP to assert that the jumplevel popped is the top ↵Gerard Goossen2009-10-311-0/+1
| | | | level jumplevel
* Add clear magic to %^H so that the HE chain is reset when you empty it.Zefram2009-08-211-4/+9
| | | | This fixes [perl #68590] : %^H not lexical enough.
* Remove all #ifdef MACOS_TRADITIONAL code in core and non-dual-life XS code.Nicholas Clark2009-04-271-5/+1
| | | | | | | | (MacOS support was removed from MakeMaker in 6.22, and merged to blead on 15th December 2004 with 5dca256ec738057dc331fb644a93eca44ad5fa14. After this point MacOS wouldn't even have been able to build the perl binary, because it would not have been able to build DynaLoader. If anyone wishes to resurrect MacOS, start by reversing this commit and the relevant part of that commit.)
* Fix MULTICALL in List-UtilAlex Vandiver2009-03-071-0/+3
| | | | | | | | Add a new stack frame; the MULTICALL code in pp_return assumes there is a clean stack, and blindly changes the top-most stack value. This change brings cop.h more in line with the "poor-man's MULTICALL" provided in ext/List-Util/multicall.h, which also always pushes a new stack.
* rxres_free() and rxres_restore() are only used in pp_ctl.c, so can beNicholas Clark2008-11-261-4/+6
| | | | | static. Macros PUSHSUBST() and POPSUBST() are only viable in PERL_CORE. p4raw-id: //depot/perl@34935
* A couple of (COP*)s that can actually be (const COP *)s.Nicholas Clark2008-10-281-2/+2
| | | p4raw-id: //depot/perl@34621
* Convert all (CV *) casts to (const CV *). Convert (XPVCV*) casts to addNicholas Clark2008-10-271-4/+4
| | | | | MUTABLE_PTR(), to validate that there is no casting away of const. p4raw-id: //depot/perl@34610
* Add a macro MUTABLE_PTR(p), which on (non-pedantic) gcc will not castNicholas Clark2008-10-271-2/+2
| | | | | | | | | | away const, returning a void *. Add MUTABLE_SV(sv) which uses this, and replace all (SV *) casts either with MUTABLE_SV(sv), or (const SV *). This probably still needs some work - assigning to SvPVX() and SvRV() is now likely to generate a casting error. The core doesn't do this. But as-is it's finding bugs that can be fixed. p4raw-id: //depot/perl@34605
* Update copyright years.Nicholas Clark2008-10-251-2/+2
| | | p4raw-id: //depot/perl@34585
* Eliminate cop_label from struct cop by storing a label as the firstNicholas Clark2008-04-071-11/+3
| | | | | | entry in the hints hash. Most statements don't have labels, so this will save memory. Not sure how much. p4raw-id: //depot/perl@33656
* Change 33653 (inevitably) missed one cop_label, because I mistook itNicholas Clark2008-04-061-1/+1
| | | | | for part of some other context-stack related struct. p4raw-id: //depot/perl@33655
* A macro used only once isn't a typing saving, and 3 macro definitionsNicholas Clark2008-04-061-1/+1
| | | | | | | never used are more obfuscation than clarification, so inline the only use of SAVECOPLABEL_FREE(), and remove the unthreaded variant and both SAVECOPLABEL()s. Exterminate! Exterminate! Exterminate! p4raw-id: //depot/perl@33654
* Micro-optimise the order of the context types. [Because I can :-)]Nicholas Clark2008-02-041-12/+19
| | | | | Here saves 72 bytes in pp_ctl.o. Small, but in the right direction. p4raw-id: //depot/perl@33232
* The third argument of PUSHEVAL was never used by anything. Ever.Nicholas Clark2008-01-281-1/+1
| | | p4raw-id: //depot/perl@33086
* The layout for struct block_loop under ithreads can be simplified.Nicholas Clark2008-01-261-17/+18
| | | | | | | | | | | | | Instead of wedging the pad offset into a void* iterdata, and always storing PL_comppad even when it isn't used, instead do this: PAD *oldcomppad; /* Also used for the GV, if targoffset is 0 */ /* This is also accesible via cx->blk_loop.my_op->op_targ */ PADOFFSET targoffset; and store the GV pointer in oldcompad. Pointers to pointers seems cleaner. This also allows us to eliminate the flag bit CXp_PADVAR. p4raw-id: //depot/perl@33081
* Investigation reveals that the work of restoring the iterator to theNicholas Clark2008-01-261-16/+2
| | | | | | | | | | | | | | | | | pad is shared between POPLOOP, using itersave, and the end of scope restore action requested by Perl_save_padsv(). In fact, the only user of SAVEt_PADSV is pp_enteriter, and it already provides enough information to allow it to perform the sv_2mortal() in POPLOOP. So make it do so. Rather than creating a new routine, use the existing routine because nothing else (at least nothing else known to Google's codesearch) uses it. But rename it just in case something we can't see is being naughty and using our private functions - they will get link errors against 5.12. All this means that itersave is now redundant. So remove it. This makes struct context 48 bytes on ILP32 platforms with 32bit IVs, down from 64 bytes in 5.10. 33% more context stack in the same memory. p4raw-id: //depot/perl@33080
View Lorry Controller Status