summaryrefslogtreecommitdiff
path: root/sv.c
Commit message (Collapse)AuthorAgeFilesLines
* Remove duplicate "the" in commentsElvin Aslanov2023-05-031-1/+1
| | | | Fix spelling on various files pertaining to core Perl.
* sv.c - don't cast SVTYPEMASK to svtype, use SvIS_FREED() instead.Yves Orton2023-03-291-5/+5
| | | | | | | svtype is an enum with 18 values. SVTYPEMASK is 31. A picky compiler (like on HPUX) will complain that casting 31 to a svtype is an error. We have SvIS_FREED() to do this properly anyway.
* report uninit variable name even in optimized casesLukas Mai2023-03-241-1/+32
| | | | | | | | | | | | | | | | | | Several cases that used to be simple assignment ops with lexical variables have been optimized in some way: - $foo = undef is now a single OP_UNDEF with special flags - $foo = ... is now a single OP_PADSV_STORE - $foo[0] = ... is now a single OP_AELEMFASTLEX_STORE This is mostly transparent to users, except for "Use of uninitialized value" warnings, which previously mentioned the name of the undefined variable, but don't do so anymore in blead. This commit teaches find_uninit_var() about the new ops, so error messages for these ops can mention variable names again. Fixes #20945.
* Mark vtbl argument in sv_unmagicext constLeon Timmermans2023-03-181-2/+2
|
* pp_ctl.c - add support for hooking require.Yves Orton2023-03-181-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | This defines a new magic hash C<%{^HOOK}> which is intended to be used for hooking keywords. It is similar to %SIG in that the values it contains are validated on set, and it is not allowed to store something in C<%{^HOOK}> that isn't supposed to be there. Hooks are expected to be coderefs (people can use currying if they really want to put an object in there, the API is deliberately simple.) The C<%{^HOOK}> hash is documented to have keys of the form "${keyword}__${phase}" where $phase is either "before" or "after" and in this initial release two hooks are supported, "require__before" and "require__after": The C<require__before> hook is called before require is executed, including any @INC hooks that might be fired. It is called with the path of the file being required, just as would be stored in %INC. The hook may alter the filename by writing to $_[0] and it may return a coderef to be executed *after* the require has completed, otherwise the return is ignored. This coderef is also called with the path of the file which was required, and it will be called regardless as to whether the require (or its dependencies) die during execution. This mechanism makes it trivial and safe to share state between the initial hook and the coderef it returns. The C<require__after> hook is similar to the C<require__before> hook however except that it is called after the require completes (successfully or not), and its return is ignored always.
* scope.c - improved RCPV support, SAVERCPV SAVEFREERCPVYves Orton2023-03-061-1/+5
| | | | | | | | | | | | | | | | | | | | | | | I misnamed some of the RCPV stuff when it was introduced. The macro which I called SAVERCPVFREE should have been called SAVERCPV, and there should also have been a SAVEFREERCPV macro. Note the naming system for these functions is a bit confusing. SAVEFREERCPV /just/ frees. SAVERCPV saves and restores (eg, its like local). SAVEFREESV is an exception. :-( and it behaves like SAVERCPV. See for instance SAVEPV or similar macros. There also should have been RCPV_REFCNT_dec() and RCPV_REFCNT_inc() macros. There was also missing documentation. This patch should fix all of these issues. Since this functionality has never been released in a production perl it should be fine to do these renames, nothing out there should use these macros yet. I noticed these oversights while fixing Scope::Upper, see also: https://github.com/Perl/perl5/issues/20861 https://rt.cpan.org/Ticket/Display.html?id=146897
* NULL sv->sv_debug_file when freedDavid Mitchell2023-02-281-1/+4
| | | | | | | | This field is used on DEBUG_LEAKING_SCALARS builds to track the file where the SV was allocated. When freeing the SV, the string was freed but the pointer was left pointing at the freed string. So NULL it out.
* Fix a bunch of memory leaks in feature 'class'Paul "LeoNerd" Evans2023-02-171-0/+6
| | | | | | | | | | | * Free the attrlist OP fragment when applying class or field attribute * Free the OP_PADxV ops we only use to get the pad index out for fieldvar declarations * Add a refcount to the `struct padname_fieldinfo` to keep track of its capture in inner closures so it can be freed at the right time * Free the class-related fields out of HvAUX * Free the actual ObjectFIELDS() array when destroying an object instance * Dup fieldinfo->paramname at sv_dup() time / free it at free time
* Ensure that sv_dup() handles new class structuresPaul "LeoNerd" Evans2023-02-131-1/+21
| | | | | | | It needs to: * clone SVt_PVOBJ instances * clone the xhv_class_* fields of an SVt_PVHV * clone the PadnameFIELDINFO structure of a padname
* Should set the SVphv_HasAUX flag explicitly; it isn't necessarily SvOOK any morePaul "LeoNerd" Evans2023-02-131-1/+1
|
* Refactor out the part of sv_dup that clones the HvAUX structure into its own ↵Paul "LeoNerd" Evans2023-02-131-64/+68
| | | | helper function
* Move the macros which wrap sv_dup_inc() into sv.hPaul "LeoNerd" Evans2023-02-131-15/+0
|
* sv.c: fix compilation with g++ <= 7Tomasz Konojacki2023-02-121-0/+1
| | | | | | | | | | | | | This compiler predates the addition of designated initializers to the C++ standard, so it's not surprising its implementation isn't fully compliant. Omitting a field in a designated initializer list causes the following error: sv.c: In function ‘void Perl_sv_upgrade(SV*, svtype)’: sv.c:1088:13: sorry, unimplemented: non-trivial designated initializers not supported
* Must call Perl_croak() on the proto_perl, not the target perl, when failing ↵Paul "LeoNerd" Evans2023-02-121-1/+1
| | | | during sv_inc()
* Perl_sv_clear: faster treatment of bodyless NVsRichard Leach2023-02-101-1/+7
| | | | | The existing fast path for SVt_NULL/SVt_IV can also be used for bodyless SVt_NVs.
* sv_clear apidoc: prefer SvREFCNT_dec to sv_free()Richard Leach2023-02-101-2/+1
| | | | | | | As per the old test, SvREFCNT_dec used to wrap sv_free, which then calls the Perl_SvREFCNT_dec function. Nowadays, SvREFCNT_dec calls Perl_SvREFCNT_dec directly, and the latter is a very small inline function. Using the macro probably should be the go-to now.
* Disallow bless hackery around classes and instancesPaul "LeoNerd" Evans2023-02-101-0/+5
|
* Create a specific SV type for object instancesPaul "LeoNerd" Evans2023-02-101-2/+37
|
* warn on redefining an array constant subTony Cook2023-02-081-1/+1
| | | | | | | | | | | | | | | | | | | The code in report_redefined_cv() would check whether the old constant and the new constant were the same SV to avoid warning on duplicate imports of the same constant. Since both gv_setref() and report_redefined_cv() used cv_const_sv() to fetch the constant SV, which returns NULL for an AV based constant sub, the check for the equivalent SV would compare NULL to NULL and report_redefined_cv() would return early. Additional checks were required further down to prevent sv_cmp() being called on AVs. The check for simple SV's allow redefinition of constant subs if the new value string compares the same as the old value. The AV check does not try to allow that. Fixed #20742
* perlapi: Move some embedding pod into proper sectionKarl Williamson2023-02-041-0/+1
| | | | These had gotten in the wrong section by things moving around in sv.c
* Correct typos as per GH 20435James E Keenan2022-12-291-10/+10
| | | | | | | | | | | | | | | | | | | In GH 20435 many typos in our C code were corrected. However, this pull request was not applied to blead and developed merge conflicts. I extracted diffs for the individual modified files and applied them with 'git apply', excepting four files where patch conflicts were reported. Those files were: handy.h locale.c regcomp.c toke.c We can handle these in a subsequent commit. Also, had to run these two programs to keep 'make test_porting' happy: $ ./perl -Ilib regen/uconfig_h.pl $ ./perl -Ilib regen/regcomp.pl regnodes.h
* sv.c - add support for HvNAMEf and HvNAMEf_QUOTEDPREFIX formatsYves Orton2022-12-221-0/+13
| | | | | | | | They are similar to SVf and SVf_QUOTEDPREFIX but take an HV * argument and use HvNAME() and related macros to extract the string. This is helpful as it makes constructing error messages from a stash (HV *) easier. It is the callers responsibility to ensure that the HV is actually a stash.
* sv.c - sv_2pv_flags: Fix double FETCH from tied overloaded scalarYves Orton2022-12-061-3/+10
| | | | | | | | | | | | | | When dealing with a tied scalar with get magic, and the FETCH method returned a blessed reference with overloading magic (with "a" magic), the tied scalar returned from the fetch was not copied prior to calling the magic function as an argument, this would then cause the get magic to be called again if the overloaded method happened to copy or otherwise use the tied scalar. The solution is to copy the reference prior to dispatching the overload call. It looks like we have been testing for the double FETCH for some time, without any good rationale, so this test merely changes things to expect the desired count.
* Extract minimum PV buffer/AV element size to common definitionsRichard Leach2022-11-211-5/+3
| | | | | | | | | | | | | | | | | In a nutshell, for a long time the minimum PV length (hardcoded in Perl_sv_grow) has been 10 bytes and the minimum AV array size (hardcoded in av_extend_guts) has been 4 elements. These numbers have been used elsewhere for consistency (e.g. Perl_sv_grow_fresh) in the past couple of development cycles. Having a standard definition, rather than hardcoding in multiple places, is more maintainable. This commit therefore introduces into perl.h: PERL_ARRAY_NEW_MIN_KEY PERL_STRLEN_NEW_MIN (Note: Subsequent commit(s) will actually change the values.)
* only fully calculate the stash (effective) name where neededTony Cook2022-11-181-14/+15
| | | | | | | | | gcc 12 was complaining that evaluating (somehekptr)->hek_key was always true in many places where HvNAME() or HvENAME() was being called in boolean context. Add new macros to check whether the names should be available and use those instead.
* scope.* - revert and rework SAVECOPWARNINGS changeYves Orton2022-11-041-0/+2
| | | | | | | | | | | | | | We can't put PL_compiling or PL_curcop on the save stack as we don't have a way to ensure they cross threads properly. This showed up as a win32 t/op/fork.t failure in the thread based fork emulation layer. This adds a new save type SAVEt_CURCOP_WARNINGS and macro SAVECURCOPWARNINGS() to complement SAVEt_COMPILER_WARNINGS and SAVECOMPILEWARNINGS(). By simply hard coding where the pointers should be restored to we side step the issue of which thread we are in. Thanks to Graham Knop for help identifying that one of my commits was responsible.
* cop.h - get rid of the STRLEN* stuff from cop_warningsYves Orton2022-11-021-1/+1
| | | | | With RCPV strings we can use the RCPV_LEN() macro, and make this logic a little less weird.
* We freed PL_splitstr in each thread, allocate it in each thread tooTony Cook2022-11-021-1/+1
| | | | | | | | I also fixed const correctness for PL_splitstr, it was initialized to " " which required it to be const, but the value is only looked at when PL_minus_F is true, so it wasn't necessary. So initialize it to NULL, and duplicate the value in threads.
* cop.h - add support for refcounted filenames in cops under threadsYves Orton2022-11-011-3/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We have a weird bifurcation of the cop logic around threads. With threads we use a char * cop_file member, without it we use a GV * and replace cop_file with cop_filegv. The GV * code refcounts filenames and more or less efficiently shares the filename amongst many opcodes. However under threads we were simplify copying the filenames into each opcode. This is because in theory opcodes created in one thread can be destroyed in another. I say in theory because as far as I know the core code does not actually do this. But we have tests that you can construct a perl, clone it, and then destroy the original, and have the copy work just fine, this means that opcodes constructed in the main thread will be destroyed in the cloned thread. This in turn means that you can't put SV derived structures into the op-tree under threads. Which is why we can not use the GV * stategy under threads. As such this code adds a new struct/type RCPV, which is a refcounted string using shared memory. This is implemented in such a way that code that previously used a char * can continue to do so, as the refcounting data is located a specific offset before the char * pointer itself. This also allows the len data to embedded "into" the PV, which allows us to expose macros to acces the length of what is in theory a null terminated string. struct rcpv { UV refcount; STRLEN len; char pv[1]; }; typedef struct rcpv RCPV; The struct is sized appropriately on creation in rcpv_new() so that the pv member contains the full string plus a null byte. It then returns a pointer to the pv member of the struct. Thus the refcount and length and embedded at a predictable offset in front of the char *, which means we do not have to change any types for members using this. We provide three operations: rcpv_new(), rcpv_copy() and rcpv_free(), which roughly correspond with newSVpv(), SvREFCNT_inc(), SvREFCNT_dec(), and a handful of macros as well. We also expose SAVERCPVFREE which is similar to SAVEGENERICSV but operates on pv's constructed with rcpv_new(). Currently I have not restricted use of this logic to threaded perls. We simply do not use it in unthreaded perls, but I see no reason we couldn't normalize the code to use this in both cases, except possibly that actually the GV case is more efficient. Note that rcpv_new() does NOT use a hash table to dedup strings. Two calls to rcpv_new() with the same arguments will produce two distinct pointers with their own refcount data. Refcounting the cop_file data was Tony Cook's idea.
* sv.c - add BODYLESS_NV fast code to sv_setXv functionsRichard Leach2022-10-221-4/+51
| | | | | | | | | | | | | | | | | For "BODYLESS" IVs and NVs, type up/downgrades can be easily done without having to call sv_upgrade. (Which will also not convert a BODYLESS NV back down to an IV, even when it is harmless to do so.) This commit adds "BODYLESS" handling to sv_setiv and sv_setnv, which previously lacked it, and "BODYLESS_NV" handling to sv_setsv_flags, which previously only special-cased "BODYLESS_IV"s. The BODYLESS_IV code in sv_setsv_flags previously did not preserve flags such as SVs_TEMP or SVs_PADTMP, which seemed like a potential latent bug. (The new BODYLESS_NV code *has* to preserve SVs_TEMP in order for all tests in t/lib/warnings to continue to pass.) Rather than clearing all flags, this commit instead does a SvOK_off(dsv) for both BODYLESS_IV and BODYLESS_NV cases.
* Some locale operations need to be done in proper threadKarl Williamson2022-10-181-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This is a step in solving #20155 The POSIX 2008 locale API introduces per-thread locales. But the previous global locale system is retained, probably for backward compatibility. The POSIX 2008 interface causes memory to be malloc'd that needs to be freed. In order to do this, the caller must first stop using that memory, by switching to another locale. perl accomplishes this during termination by switching to the global locale, which is always available and doesn't need to be freed. Perl has long assumed that all that was needed to switch threads was to change out tTHX. That's because that structure was intended to hold all the information for a given thread. But it turns out that this doesn't work when some library independently holds information about the thread's state. And there are now some libraries that do that. What was happening in this case was that perl thought that it was sufficient to switch tTHX to change to a different thread in order to do the freeing of memory, and then used the POSIX 2008 function to change to the global locale so that the memory could be safely freed. But the POSIX 2008 function doesn't care about tTHX, and actually was typically operating on a different thread, and so changed that thread to the global locale instead of the intended thread. Often that was the top-level thread, thread 0. That caused whatever thread it was to no longer be in the expected locale, and to no longer be thread-safe with regards to localess, This commit causes locale_term(), which has always been called from the actual terminating thread that POSIX 2008 knows about, to change to the global thread and free the memory. It also creates a new per-interpreter variable that effectively maps the tTHX thread to the associated POSIX 2008 memory. During perl_destruct(), it frees the memory this variable points to, instead of blindly assuming the memory to free is the current tTHX thread's. This fixes the symptoms associtated with #20155, but doesn't solve the whole problem. In general, a library that has independent thread status needs to be updated to the new thread when Perl changes threads using tTHX. Future commits will do this.
* locale: Create special variable to hold current LC_ALLKarl Williamson2022-10-181-0/+3
| | | | | | | | | | | | Some configurations require us to store the current locale for each category. Prior to this commit, this was done in the array PL_curlocales, with the entry for LC_ALL being in the highest element. Future commits will need just the value for LC_ALL in some other configurations, without needing the rest of the array. This commit splits off the LC_ALL element into its own per-interpreter variable to accommodate those. It always had to have special handling anyway beyond the rest of the array elements,
* sv.c: Set phase to CONSTRUCT on interpreter being clonedKarl Williamson2022-10-181-0/+1
| | | | So far this hadn't been an issue, but it will be for a future commit.
* sv.c: Clone interpreter with locale set to CKarl Williamson2022-10-181-19/+20
| | | | | | A thread is supposed to start with the locale set to the C locale. We were duping the parent values, and later overriding. Better to set to C from the beginning.
* sv.c: Move some code; consolidateKarl Williamson2022-10-181-25/+18
| | | | | | This moves the code for cloning the locale variables to a single place, consolidating some that had the same cpp directives. It showed that one variable was cloned twice; the redundant one is now removed.
* vnewSVpvf - inline & reduce call to sv_vsetpvfnRichard Leach2022-09-181-2/+3
| | | | | | | | | | sv_vsetpvfn does just two things: * SvPVCLEAR on the sv passed in * calls sv_vcatpvfn_flags Since the SV passed into sv_vsetpvfn by vnewSVpvf is a brand new SV, it is more efficient to inline sv_vsetpvfn and change the SvPVCLEAR to SvPVCLEAR_FRESH.
* sv_setsv_cow - directly create desired SVt_COW SVRichard Leach2022-09-171-2/+2
| | | | | It's more efficient to directly create the desired SV type than create a SVt_NULL and then call sv_upgrade to get the desired type.
* Perl_sv_setsv_flags: return after invlist_clone caseRichard Leach2022-09-141-2/+4
| | | | | | | | | | | | | | | | | | | | | | | | In the "first_ switch" statement in Perl_sv_setsv_flags, there's this case: case SVt_INVLIST: invlist_clone(ssv, dsv); break; When the INVLIST code was added, it was unclear whether the logic following that switch needed to apply to INVLISTs or not. Early returns were also seen as risky from a future maintenance perspective. As a precaution, this case was written to `break` rather than `return`. With later analysis though, it seems like the code below this switch does not apply to INVLISTs, apart from one unintended interaction. * invlist_clone() - in regcomp.c - copies the PV buffer of ssv to dsv. * However, whenever an invlist is initialized, SvPOK_on is called on its SV; according to the nearby comment "/* This allows B to extract the PV */". * Because the switch case does not return, the later `if (sflags & SVp_POK)` branch will try to swipe/cow/copy the PV buffer of ssv to dsv. This is an unnecessary double copy. On that analysis, this commit changes the break into a return.
* sv.c: Comments white-space onlyKarl Williamson2022-09-091-6/+6
|
* Use general locale mutex for numeric operationsKarl Williamson2022-09-091-3/+0
| | | | | | | | | | | | | | | | This commit removes the separate mutex for locking locale-related numeric operations on threaded perls; instead using the general locale one. The previous commit made that a general semaphore, so now suitable for use for this purpose as well. This means that the locale can be locked for the duration of some sprintf operations, longer than before this commit. But on most modern platforms, thread-safe locales cause this lock to expand just to a no-op; so there is no effect on these. And on the impacted platforms, one is not supposed to be using locales and threads in combination, as races can occur. This lock is used on those perls to keep Perl's manipulation of LC_NUMERIC thread-safe. And for those there is also no effect, as they already lock around those sprintf's.
* Make the locale mutex a general semaphoreKarl Williamson2022-09-091-0/+5
| | | | | Future commits will use this new capability, and in Configurations where no locale locking is currently necessary.
* locale.c: Move #define to perl.h; use it elsewhereKarl Williamson2022-09-091-1/+1
| | | | | Rather than recalculate this combined conditional, do it once in perl.h.
* intrpvar.h,sv.c,perl.c: Remove unnecessary cpp conditionKarl Williamson2022-09-091-3/+1
| | | | | | This conditional dates from when the rest of the conditions used 'HAS_foo' (from config.h) instead of USE_foo, which takes more things into account.
* locale.c: Rmv no longer used code; UTF8ness cacheKarl Williamson2022-09-021-1/+0
| | | | | | | | | | | | | What these functions do has been subsumed by code introduced in previous commits, and in a more straight forward manner. Also removed in this commit is the cache of the knowing what locales are UTF-8 or not. This data is now cheaper to calculate when needed, and there is now a single entry cache, so I don't think the complexity warrants keeping it. It could be added back if necessary, split off from the remainder of this commit.
* op.c - Restrict nested eval/BEGIN blocks to a user controllable maximumYves Orton2022-09-021-0/+1
| | | | | | | | | | | | | | | Nested BEGIN blocks can cause us to segfault by exhausting the C stack. Eg: perl -le'sub f { eval "BEGIN { f() }" } f()' will segfault. This adds a new interpreter var PL_eval_begin_nest_depth to keep track of how many layer of eval/BEGIN we have seen, and a new reserved variable called ${^MAX_NESTED_EVAL_BEGIN_BLOCKS} which can be used to raise or lower the limit. When set to 0 it blocks BEGIN entirely, which might be useful from time to time. This fixes https://github.com/Perl/perl5/issues/20176
* Change internal function name to be standards compliantKarl Williamson2022-09-011-1/+1
| | | | | This is in preparation for working on it; the new name, mem_collxfrm_ is in compliance with the C Standard; the old was not.
* locale.c: Save underlying radix characterKarl Williamson2022-09-011-0/+1
| | | | | | | | | | | When changing locales the new decimal point needs to be calculated. This commit creates a new per-interpreter variable to save that calculation, so it only has to be done when a new locale is set; prior to this commit it was recalculated each time it was needed. The calculation is still performed twice when the new locale is switched into. But the redundant calculation will be removed in a couple of commits hence.
* locale.c: Cache the current LC_CTYPE locale nameKarl Williamson2022-08-311-0/+1
| | | | | | | | This is now used as a cache of length 1 to avoid having to lookup up the UTF-8ness as often. This commit also skips doing S_newctype() if the new boss is the same as the old
* sv.c - add a _QUOTEDPREFIX version of SVf, UTF8f, and HEKf for use in error ↵Yves Orton2022-08-251-9/+93
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | messages. These new formats are intended to be used in error messages where we want to show the contents of a string without any possible hidden characters not rendering in the error message, and where it would be unreasonable to show every character of the string if it is very long. A good example would be when we want to say that a class name is illegal. Consider: "Foo\0"->thing() should not throw an error message about "Foo" being missing, the fact there is a null in there should be visible to the developer. Similarly if we had ("x" x 1000_000)->thing() we also do not want to throw a 1MB error message as it is generally just unhelpful, a class name that long is almost certainly a mistake. Currently this patch restricts it to showing 256 characters, the first 128 followed by an ellipses followed by the last 128 characters, but the docs are such that we can change that if we wish, I suspect something like 100 would be more reasonable. You can override the define PERL_QUOTEDPREFIX_LEN to a longer value in Configure if you wish. Example usage: other= newSVpvs("Some\0::Thing\n"); sv_catpvf(msg_sv,"%" SVf_QUOTEDPREFIX, SVfARG(other)); Should append "Some\0::Thing\n" to the msg_sv. If it were very long it would have ellipses infixed. The class name "x" x 1_000_000 would show Can't locate object method "non_existent_method" via \ package "x[repeated 128 times]"..."x[repeated 128 times]" \ (perhaps you forgot to load \ "x[repeated 128 times]"..."x[repeated 128 times]"?) at -e line 1. (but obviously as one line with the literal text of the class instead of "[repeated 128 times]") This patch changes a variety of error messages that used to output the full string always. I haven't changed every place that this could happen yet, just the main ones related to method calls, subroutine names and the like.
* Define a CvREFCOUNTED_ANYSV flagPaul "LeoNerd" Evans2022-08-161-0/+3
| | | | | | | | If this flag is set, then the CvXSUBANY(cv).any_sv pointer will have its reference count decremented when the CV itself is freed. This is useful for XS extensions that wish to store extra data in here. Without this flag, such extensions have to resort to using magic with a 'free' function to perform this work.
View Lorry Controller Status