Re: [PATCH] Cleanup of the regexp API

[p5sagit/p5-mst-13.2.git] / pod / perlreapi.pod

=head1 NAME

perlreapi - perl regular expression plugin interface

=head1 DESCRIPTION

As of Perl 5.9.5 there is a new interface for using other regexp engines than
the default one.  Each engine is supposed to provide access to a constant
structure of the following format:

    typedef struct regexp_engine {
        REGEXP* (*comp) (pTHX_ const SV * const pattern, const U32 flags);
        I32     (*exec) (pTHX_ REGEXP * const rx, char* stringarg, char* strend,
                    char* strbeg, I32 minend, SV* screamer,
                    void* data, U32 flags);
        char*   (*intuit) (pTHX_ REGEXP * const rx, SV *sv, char *strpos,
                    char *strend, U32 flags,
                    struct re_scream_pos_data_s *data);
        SV*     (*checkstr) (pTHX_ REGEXP * const rx);
        void    (*free) (pTHX_ REGEXP * const rx);
        void    (*numbered_buff_get) (pTHX_ REGEXP * const rx,
                    const I32 paren, SV * const usesv);
        SV*     (*named_buff_get)(pTHX_ REGEXP * const rx, SV * const namesv,
                    const U32 flags);
        SV*     (*qr_package)(pTHX_ REGEXP * const rx);
    #ifdef USE_ITHREADS
        void*   (*dupe) (pTHX_ REGEXP * const rx, CLONE_PARAMS *param);
    #endif
    } regexp_engine;

When a regexp is compiled, its C<engine> field is then set to point at
the appropriate structure so that when it needs to be used Perl can find
the right routines to do so.

In order to install a new regexp handler, C<$^H{regcomp}> is set
to an integer which (when casted appropriately) resolves to one of these
structures. When compiling, the C<comp> method is executed, and the
resulting regexp structure's engine field is expected to point back at
the same structure.

The pTHX_ symbol in the definition is a macro used by perl under threading
to provide an extra argument to the routine holding a pointer back to
the interpreter that is executing the regexp. So under threading all
routines get an extra argument.

The routines are as follows:

=head2 comp

    REGEXP* comp(pTHX_ const SV * const pattern, const U32 flags);

Compile the pattern stored in C<pattern> using the given C<flags> and
return a pointer to a prepared C<REGEXP> structure that can perform
the match. See L</The REGEXP structure> below for an explanation of
the individual fields in the REGEXP struct.

The C<pattern> parameter is the scalar that was used as the
pattern. previous versions of perl would pass two C<char*> indicating
the start and end of the stringifed pattern, the following snippet can
be used to get the old parameters:

    STRLEN plen;
    char*  exp = SvPV(pattern, plen);
    char* xend = exp + plen;

Since any scalar can be passed as a pattern it's possible to implement
an engine that does something with an array (C<< "ook" =~ [ qw/ eek
hlagh / ] >>) or with the non-stringified form of a compiled regular
expression (C<< "ook" =~ qr/eek/ >>). perl's own engine will always
stringify everything using the snippet above but that doesn't mean
other engines have to.

The C<flags> paramater is a bitfield which indicates which of the
C<msixk> flags the regex was compiled with. In addition it contains
info about whether C<use locale> is in effect and optimization info
for C<split>. A regex engine might want to use the same split
optimizations with a different syntax, for instance a Perl6 engine
would treat C<split /^^/> equivalently to perl's C<split /^/>, see
L<split documentation|perlfunc> and the relevant code in C<pp_split>
in F<pp.c> to find out whether your engine should be setting these.

The C<eogc> flags are stripped out before being passed to the comp
routine. The regex engine does not need to know whether any of these
are set as those flags should only affect what perl does with the
pattern and its match variables, not how it gets compiled & executed.

=over 4

=item RXf_SKIPWHITE

C<split ' '> or C<split> with no arguments (which really means
C<split(' ', $_> see L<split|perlfunc>).

=item RXf_START_ONLY

Set if the pattern is C</^/> (C<<r->prelen == 1 && r->precomp[0] ==
'^'>>). Will be used by the C<split> operator to split the given
string on C<\n> (even under C</^/s>, see L<split|perlfunc>).

=item RXf_WHITE

Set if the pattern is exactly C</\s+/> and used by C<split>, the
definition of whitespace varies depending on whether RXf_UTF8 or
RXf_PMf_LOCALE is set.

=item RXf_PMf_LOCALE

Makes C<split> use the locale dependant definition of whitespace under C<use
locale> when RXf_SKIPWHITE or RXf_WHITE is in effect. Under ASCII whitespace is
defined as per L<isSPACE|perlapi/ISSPACE>, and by the internal macros
C<is_utf8_space> under UTF-8 and C<isSPACE_LC> under C<use locale>.

=item RXf_PMf_MULTILINE

The C</m> flag, this ends up being passed to C<Perl_fbm_instr> by
C<pp_split> regardless of the engine.

=item RXf_PMf_SINGLELINE

The C</s> flag. Guaranteed not to be used outside the regex engine.

=item RXf_PMf_FOLD

The C</i> flag. Guaranteed not to be used outside the regex engine.

=item RXf_PMf_EXTENDED

The C</x> flag. Guaranteed not to be used outside the regex
engine. However if present on a regex C<#> comments will be stripped
by the tokenizer regardless of the engine currently in use.

=item RXf_PMf_KEEPCOPY

The C</p> flag.

=item RXf_UTF8

Set if the pattern is L<SvUTF8()|perlapi/SvUTF8>, set by Perl_pmruntime.

=back

In general these flags should be preserved in regex->extflags after
compilation, although it is possible the regex includes constructs
that changes them. The perl engine for instance may upgrade non-utf8
strings to utf8 if the pattern includes constructs such as C<\x{...}>
that can only match unicode values. RXf_SKIPWHITE should always be
preserved verbatim in regex->extflags.

=head2 exec

    I32 exec(pTHX_ REGEXP * const rx,
             char *stringarg, char* strend, char* strbeg,
             I32 minend, SV* screamer,
             void* data, U32 flags);

Execute a regexp.

=head2 intuit

    char* intuit(pTHX_ REGEXP * const rx,
                  SV *sv, char *strpos, char *strend,
                  const U32 flags, struct re_scream_pos_data_s *data);

Find the start position where a regex match should be attempted,
or possibly whether the regex engine should not be run because the
pattern can't match. This is called as appropriate by the core
depending on the values of the extflags member of the regexp
structure.

=head2 checkstr

    SV*	checkstr(pTHX_ REGEXP * const rx);

Return a SV containing a string that must appear in the pattern. Used
by C<split> for optimising matches.

=head2 free

    void free(pTHX_ REGEXP * const rx);

Called by perl when it is freeing a regexp pattern so that the engine
can release any resources pointed to by the C<pprivate> member of the
regexp structure. This is only responsible for freeing private data;
perl will handle releasing anything else contained in the regexp structure.

=head2 numbered_buff_get

    void numbered_buff_get(pTHX_ REGEXP * const rx, const I32 paren,
                           SV * const usesv);

Called to get the value of C<$`>, C<$'>, C<$&> (and their named
equivalents, see L<perlvar>) and the numbered capture buffers (C<$1>,
C<$2>, ...).

The C<paren> paramater will be C<-2> for C<$`>, C<-1> for C<$'>, C<0>
for C<$&>, C<1> for C<$1> and so forth.

C<usesv> should be set to the scalar to return, the scalar is passed
as an argument rather than being returned from the function because
when it's called perl already has a scalar to store the value,
creating another one would be redundant. The scalar can be set with
C<sv_setsv>, C<sv_setpvn> and friends, see L<perlapi>.

This callback is where perl untaints its own capture variables under
taint mode (see L<perlsec>). See the C<Perl_reg_numbered_buff_get>
function in F<regcomp.c> for how to untaint capture variables if
that's something you'd like your engine to do as well.

=head2 named_buff_get

    SV* named_buff_get(pTHX_ REGEXP * const rx, SV * const namesv,
                       const U32 flags);

Called to get the value of key in the C<%+> and C<%-> hashes,
C<namesv> is the hash key being requested and if C<flags & 1> is true
C<%-> is being requested (and C<%+> if it's not).

=head2 qr_package

    SV* qr_package(pTHX_ REGEXP * const rx);

The package the qr// magic object is blessed into (as seen by C<ref
qr//>). It is recommended that engines change this to their package
name for identification regardless of whether they implement methods
on the object.

A callback implementation might be:

    SV*
    Example_reg_qr_package(pTHX_ REGEXP * const rx)
    {
    	PERL_UNUSED_ARG(rx);
    	return newSVpvs("re::engine::Example");
    }

Any method calls on an object created with C<qr//> will be dispatched to the
package as a normal object.

    use re::engine::Example;
    my $re = qr//;
    $re->meth; # dispatched to re::engine::Example::meth()

To retrieve the C<REGEXP> object from the scalar in an XS function use the
following snippet:

    void meth(SV * rv)
    PPCODE:
        MAGIC  * mg;
        REGEXP * re;

        if (SvMAGICAL(sv))
            mg_get(sv);
        if (SvROK(sv) &&
            (sv = (SV*)SvRV(sv)) &&            /* assignment deliberate */
            SvTYPE(sv) == SVt_PVMG &&
            (mg = mg_find(sv, PERL_MAGIC_qr))) /* assignment deliberate */
        {
            re = (REGEXP *)mg->mg_obj; 
        }

Or use the (CURRENTLY UNDOCUMENETED!) C<Perl_get_re_arg> function:

    void meth(SV * rv)
    PPCODE:
        const REGEXP * const re = (REGEXP *)Perl_get_re_arg( aTHX_ rv, 0, NULL );

=head2 dupe

    void* dupe(pTHX_ REGEXP * const rx, CLONE_PARAMS *param);

On threaded builds a regexp may need to be duplicated so that the pattern
can be used by mutiple threads. This routine is expected to handle the
duplication of any private data pointed to by the C<pprivate> member of
the regexp structure.  It will be called with the preconstructed new
regexp structure as an argument, the C<pprivate> member will point at
the B<old> private structue, and it is this routine's responsibility to
construct a copy and return a pointer to it (which perl will then use to
overwrite the field as passed to this routine.)

This allows the engine to dupe its private data but also if necessary
modify the final structure if it really must.

On unthreaded builds this field doesn't exist.

=head1 The REGEXP structure

The REGEXP struct is defined in F<regexp.h>. All regex engines must be able to
correctly build such a structure in their L</comp> routine.

The REGEXP structure contains all the data that perl needs to be aware of
to properly work with the regular expression. It includes data about
optimisations that perl can use to determine if the regex engine should
really be used, and various other control info that is needed to properly
execute patterns in various contexts such as is the pattern anchored in
some way, or what flags were used during the compile, or whether the
program contains special constructs that perl needs to be aware of.

In addition it contains two fields that are intended for the private use
of the regex engine that compiled the pattern. These are the C<intflags>
and pprivate members. The C<pprivate> is a void pointer to an arbitrary
structure whose use and management is the responsibility of the compiling
engine. perl will never modify either of these values.

    typedef struct regexp {
        /* what engine created this regexp? */
        const struct regexp_engine* engine;

        /* what re is this a lightweight copy of? */
        struct regexp* mother_re;

        /* Information about the match that the perl core uses to manage things */
        U32 extflags;   /* Flags used both externally and internally */
        I32 minlen;     /* mininum possible length of string to match */
        I32 minlenret;  /* mininum possible length of $& */
        U32 gofs;       /* chars left of pos that we search from */

        /* substring data about strings that must appear
           in the final match, used for optimisations */
        struct reg_substr_data *substrs;

        U32 nparens;  /* number of capture buffers */

        /* private engine specific data */
        U32 intflags;   /* Engine Specific Internal flags */
        void *pprivate; /* Data private to the regex engine which 
                           created this object. */

        /* Data about the last/current match. These are modified during matching*/
        U32 lastparen;            /* last open paren matched */
        U32 lastcloseparen;       /* last close paren matched */
        regexp_paren_pair *swap;  /* Swap copy of *offs */
        regexp_paren_pair *offs;  /* Array of offsets for (@-) and (@+) */

        char *subbeg;  /* saved or original string so \digit works forever. */
        SV_SAVED_COPY  /* If non-NULL, SV which is COW from original */
        I32 sublen;    /* Length of string pointed by subbeg */

        /* Information about the match that isn't often used */
        I32 prelen;           /* length of precomp */
        const char *precomp;  /* pre-compilation regular expression */

        /* wrapped can't be const char*, as it is returned by sv_2pv_flags */
        char *wrapped;  /* wrapped version of the pattern */
        I32 wraplen;    /* length of wrapped */

        I32 seen_evals;   /* number of eval groups in the pattern - for security checks */
        HV *paren_names;  /* Optional hash of paren names */

        /* Refcount of this regexp */
        I32 refcnt;             /* Refcount of this regexp */
    } regexp;

The fields are discussed in more detail below:

=over 4

=item C<engine>

This field points at a regexp_engine structure which contains pointers
to the subroutines that are to be used for performing a match. It
is the compiling routine's responsibility to populate this field before
returning the regexp object.

Internally this is set to C<NULL> unless a custom engine is specified in
C<$^H{regcomp}>, perl's own set of callbacks can be accessed in the struct
pointed to by C<RE_ENGINE_PTR>.

=item C<mother_re>

TODO, see L<http://www.mail-archive.com/perl5-changes@perl.org/msg17328.html>

=item C<extflags>

This will be used by perl to see what flags the regexp was compiled with, this
will normally be set to the value of the flags parameter on L</comp>.

=item C<minlen> C<minlenret>

The minimum string length required for the pattern to match.  This is used to
prune the search space by not bothering to match any closer to the end of a
string than would allow a match. For instance there is no point in even
starting the regex engine if the minlen is 10 but the string is only 5
characters long. There is no way that the pattern can match.

C<minlenret> is the minimum length of the string that would be found
in $& after a match.

The difference between C<minlen> and C<minlenret> can be seen in the
following pattern:

    /ns(?=\d)/

where the C<minlen> would be 3 but C<minlenret> would only be 2 as the \d is
required to match but is not actually included in the matched content. This
distinction is particularly important as the substitution logic uses the
C<minlenret> to tell whether it can do in-place substition which can result in
considerable speedup.

=item C<gofs>

Left offset from pos() to start match at.

=item C<substrs>

TODO: document

=item C<nparens>, C<lasparen>, and C<lastcloseparen>

These fields are used to keep track of how many paren groups could be matched
in the pattern, which was the last open paren to be entered, and which was
the last close paren to be entered.

=item C<intflags>

The engine's private copy of the flags the pattern was compiled with. Usually
this is the same as C<extflags> unless the engine chose to modify one of them

=item C<pprivate>

A void* pointing to an engine-defined data structure. The perl engine uses the
C<regexp_internal> structure (see L<perlreguts/Base Structures>) but a custom
engine should use something else.

=item C<swap>

TODO: document

=item C<offs>

A C<regexp_paren_pair> structure which defines offsets into the string being
matched which correspond to the C<$&> and C<$1>, C<$2> etc. captures, the
C<regexp_paren_pair> struct is defined as follows:

    typedef struct regexp_paren_pair {
        I32 start;
        I32 end;
    } regexp_paren_pair;

If C<< ->offs[num].start >> or C<< ->offs[num].end >> is C<-1> then that
capture buffer did not match. C<< ->offs[0].start/end >> represents C<$&> (or
C<${^MATCH> under C<//p>) and C<< ->offs[paren].end >> matches C<$$paren> where
C<$paren >= 1>.

=item C<precomp> C<prelen>

Used for debugging purposes. C<precomp> holds a copy of the pattern
that was compiled and C<prelen> its length.

=item C<paren_names>

This is a hash used internally to track named capture buffers and their
offsets. The keys are the names of the buffers the values are dualvars,
with the IV slot holding the number of buffers with the given name and the
pv being an embedded array of I32.  The values may also be contained
independently in the data array in cases where named backreferences are
used.

=item C<reg_substr_data>

Holds information on the longest string that must occur at a fixed
offset from the start of the pattern, and the longest string that must
occur at a floating offset from the start of the pattern. Used to do
Fast-Boyer-Moore searches on the string to find out if its worth using
the regex engine at all, and if so where in the string to search.

=item C<subbeg> C<sublen> C<saved_copy>

    #define SAVEPVN(p,n) ((p) ? savepvn(p,n) : NULL)
    if (RX_MATCH_COPIED(ret))
        ret->subbeg = SAVEPVN(ret->subbeg, ret->sublen);
    else
        ret->subbeg = NULL;

C<PL_sawampersand || rx->extflags & RXf_PMf_KEEPCOPY>

These are used during execution phase for managing search and replace
patterns.

=item C<wrapped> C<wraplen>

Stores the string C<qr//> stringifies to, for example C<(?-xism:eek)>
in the case of C<qr/eek/>.

When using a custom engine that doesn't support the C<(?:)> construct for
inline modifiers it's best to have C<qr//> stringify to the supplied pattern,
note that this will create invalid patterns in cases such as:

    my $x = qr/a|b/;  # "a|b"
    my $y = qr/c/;    # "c"
    my $z = qr/$x$y/; # "a|bc"

There's no solution for such problems other than making the custom engine
understand some for of inline modifiers.

The C<Perl_reg_stringify> in F<regcomp.c> does the stringification work.

=item C<seen_evals>

This stores the number of eval groups in the pattern. This is used for security
purposes when embedding compiled regexes into larger patterns with C<qr//>.

=item C<refcnt>

The number of times the structure is referenced. When this falls to 0 the
regexp is automatically freed by a call to pregfree. This should be set to 1 in
each engine's L</comp> routine.

=back

=head2 De-allocation and Cloning

Any patch that adds data items to the REGEXP struct will need to include
changes to F<sv.c> (C<Perl_re_dup()>) and F<regcomp.c> (C<pregfree()>). This
involves freeing or cloning items in the regexp's data array based on the data
item's type.

=head1 HISTORY

Originally part of L<perlreguts>.

=head1 AUTHORS

Originally written by Yves Orton, expanded by E<AElig>var ArnfjE<ouml>rE<eth>
Bjarmason.

=head1 LICENSE

Copyright 2006 Yves Orton and 2007 E<AElig>var ArnfjE<ouml>rE<eth> Bjarmason.

This program is free software; you can redistribute it and/or modify it under
the same terms as Perl itself.

=cut