Mercurial > octave
view liboctave/util/lo-regexp.cc @ 31605:e88a07dec498 stable
maint: Use macros to begin/end C++ namespaces.
* oct-conf-post-public.in.h: Define two macros (OCTAVE_BEGIN_NAMESPACE,
OCTAVE_END_NAMESPACE) that can be used to start/end a namespace.
* mk-opts.pl, build-env.h, build-env.in.cc, __betainc__.cc, __contourc__.cc,
__dsearchn__.cc, __eigs__.cc, __expint__.cc, __ftp__.cc, __gammainc__.cc,
__ichol__.cc, __ilu__.cc, __isprimelarge__.cc, __lin_interpn__.cc,
__magick_read__.cc, __pchip_deriv__.cc, __qp__.cc, amd.cc, auto-shlib.cc,
auto-shlib.h, balance.cc, base-text-renderer.cc, base-text-renderer.h,
besselj.cc, bitfcns.cc, bsxfun.cc, c-file-ptr-stream.cc, c-file-ptr-stream.h,
call-stack.cc, call-stack.h, ccolamd.cc, cellfun.cc, chol.cc, colamd.cc,
colloc.cc, conv2.cc, daspk.cc, dasrt.cc, dassl.cc, data.cc, data.h, debug.cc,
defaults.cc, defaults.h, defun-int.h, defun.cc, det.cc, dirfns.cc, display.cc,
display.h, dlmread.cc, dmperm.cc, dot.cc, dynamic-ld.cc, dynamic-ld.h, eig.cc,
ellipj.cc, environment.cc, environment.h, error.cc, error.h, errwarn.h,
event-manager.cc, event-manager.h, event-queue.cc, event-queue.h, fcn-info.cc,
fcn-info.h, fft.cc, fft2.cc, fftn.cc, file-io.cc, filter.cc, find.cc,
ft-text-renderer.cc, ft-text-renderer.h, gcd.cc, getgrent.cc, getpwent.cc,
getrusage.cc, givens.cc, gl-render.cc, gl-render.h, gl2ps-print.cc,
gl2ps-print.h, graphics-toolkit.cc, graphics-toolkit.h, graphics.cc,
graphics.in.h, gsvd.cc, gtk-manager.cc, gtk-manager.h, hash.cc, help.cc,
help.h, hess.cc, hex2num.cc, hook-fcn.cc, hook-fcn.h, input.cc, input.h,
interpreter-private.cc, interpreter-private.h, interpreter.cc, interpreter.h,
inv.cc, jsondecode.cc, jsonencode.cc, kron.cc, latex-text-renderer.cc,
latex-text-renderer.h, load-path.cc, load-path.h, load-save.cc, load-save.h,
lookup.cc, ls-ascii-helper.cc, ls-ascii-helper.h, ls-oct-text.cc, ls-utils.cc,
ls-utils.h, lsode.cc, lu.cc, mappers.cc, matrix_type.cc, max.cc, mex-private.h,
mex.cc, mgorth.cc, nproc.cc, oct-fstrm.cc, oct-fstrm.h, oct-hdf5-types.cc,
oct-hdf5-types.h, oct-hist.cc, oct-hist.h, oct-iostrm.cc, oct-iostrm.h,
oct-opengl.h, oct-prcstrm.cc, oct-prcstrm.h, oct-procbuf.cc, oct-procbuf.h,
oct-process.cc, oct-process.h, oct-stdstrm.h, oct-stream.cc, oct-stream.h,
oct-strstrm.cc, oct-strstrm.h, oct-tex-lexer.in.ll, oct-tex-parser.yy,
ordqz.cc, ordschur.cc, pager.cc, pager.h, pinv.cc, pow2.cc, pr-flt-fmt.cc,
pr-output.cc, procstream.cc, procstream.h, psi.cc, qr.cc, quad.cc, quadcc.cc,
qz.cc, rand.cc, rcond.cc, regexp.cc, schur.cc, settings.cc, settings.h,
sighandlers.cc, sighandlers.h, sparse-xdiv.cc, sparse-xdiv.h, sparse-xpow.cc,
sparse-xpow.h, sparse.cc, spparms.cc, sqrtm.cc, stack-frame.cc, stack-frame.h,
stream-euler.cc, strfind.cc, strfns.cc, sub2ind.cc, svd.cc, sylvester.cc,
symbfact.cc, syminfo.cc, syminfo.h, symrcm.cc, symrec.cc, symrec.h,
symscope.cc, symscope.h, symtab.cc, symtab.h, syscalls.cc, sysdep.cc, sysdep.h,
text-engine.cc, text-engine.h, text-renderer.cc, text-renderer.h, time.cc,
toplev.cc, tril.cc, tsearch.cc, typecast.cc, url-handle-manager.cc,
url-handle-manager.h, urlwrite.cc, utils.cc, utils.h, variables.cc,
variables.h, xdiv.cc, xdiv.h, xnorm.cc, xnorm.h, xpow.cc, xpow.h,
__delaunayn__.cc, __fltk_uigetfile__.cc, __glpk__.cc, __init_fltk__.cc,
__init_gnuplot__.cc, __ode15__.cc, __voronoi__.cc, audiodevinfo.cc,
audioread.cc, convhulln.cc, fftw.cc, gzip.cc, mk-build-env-features.sh,
mk-builtins.pl, cdef-class.cc, cdef-class.h, cdef-fwd.h, cdef-manager.cc,
cdef-manager.h, cdef-method.cc, cdef-method.h, cdef-object.cc, cdef-object.h,
cdef-package.cc, cdef-package.h, cdef-property.cc, cdef-property.h,
cdef-utils.cc, cdef-utils.h, ov-base.cc, ov-base.h, ov-bool-mat.cc,
ov-builtin.h, ov-cell.cc, ov-class.cc, ov-class.h, ov-classdef.cc,
ov-classdef.h, ov-complex.cc, ov-fcn-handle.cc, ov-fcn-handle.h, ov-fcn.h,
ov-java.cc, ov-java.h, ov-mex-fcn.h, ov-null-mat.cc, ov-oncleanup.cc,
ov-struct.cc, ov-typeinfo.cc, ov-typeinfo.h, ov-usr-fcn.cc, ov-usr-fcn.h,
ov.cc, ov.h, octave.cc, octave.h, mk-ops.sh, op-b-b.cc, op-b-bm.cc,
op-b-sbm.cc, op-bm-b.cc, op-bm-bm.cc, op-bm-sbm.cc, op-cdm-cdm.cc, op-cell.cc,
op-chm.cc, op-class.cc, op-cm-cm.cc, op-cm-cs.cc, op-cm-m.cc, op-cm-s.cc,
op-cm-scm.cc, op-cm-sm.cc, op-cs-cm.cc, op-cs-cs.cc, op-cs-m.cc, op-cs-s.cc,
op-cs-scm.cc, op-cs-sm.cc, op-dm-dm.cc, op-dm-scm.cc, op-dm-sm.cc,
op-dm-template.cc, op-dms-template.cc, op-fcdm-fcdm.cc, op-fcm-fcm.cc,
op-fcm-fcs.cc, op-fcm-fm.cc, op-fcm-fs.cc, op-fcn.cc, op-fcs-fcm.cc,
op-fcs-fcs.cc, op-fcs-fm.cc, op-fcs-fs.cc, op-fdm-fdm.cc, op-fm-fcm.cc,
op-fm-fcs.cc, op-fm-fm.cc, op-fm-fs.cc, op-fs-fcm.cc, op-fs-fcs.cc,
op-fs-fm.cc, op-fs-fs.cc, op-i16-i16.cc, op-i32-i32.cc, op-i64-i64.cc,
op-i8-i8.cc, op-int-concat.cc, op-m-cm.cc, op-m-cs.cc, op-m-m.cc, op-m-s.cc,
op-m-scm.cc, op-m-sm.cc, op-mi.cc, op-pm-pm.cc, op-pm-scm.cc, op-pm-sm.cc,
op-pm-template.cc, op-range.cc, op-s-cm.cc, op-s-cs.cc, op-s-m.cc, op-s-s.cc,
op-s-scm.cc, op-s-sm.cc, op-sbm-b.cc, op-sbm-bm.cc, op-sbm-sbm.cc,
op-scm-cm.cc, op-scm-cs.cc, op-scm-m.cc, op-scm-s.cc, op-scm-scm.cc,
op-scm-sm.cc, op-sm-cm.cc, op-sm-cs.cc, op-sm-m.cc, op-sm-s.cc, op-sm-scm.cc,
op-sm-sm.cc, op-str-m.cc, op-str-s.cc, op-str-str.cc, op-struct.cc,
op-ui16-ui16.cc, op-ui32-ui32.cc, op-ui64-ui64.cc, op-ui8-ui8.cc, ops.h,
anon-fcn-validator.cc, anon-fcn-validator.h, bp-table.cc, bp-table.h,
comment-list.cc, comment-list.h, filepos.h, lex.h, lex.ll, oct-lvalue.cc,
oct-lvalue.h, oct-parse.yy, parse.h, profiler.cc, profiler.h,
pt-anon-scopes.cc, pt-anon-scopes.h, pt-arg-list.cc, pt-arg-list.h,
pt-args-block.cc, pt-args-block.h, pt-array-list.cc, pt-array-list.h,
pt-assign.cc, pt-assign.h, pt-binop.cc, pt-binop.h, pt-bp.cc, pt-bp.h,
pt-cbinop.cc, pt-cbinop.h, pt-cell.cc, pt-cell.h, pt-check.cc, pt-check.h,
pt-classdef.cc, pt-classdef.h, pt-cmd.h, pt-colon.cc, pt-colon.h, pt-const.cc,
pt-const.h, pt-decl.cc, pt-decl.h, pt-eval.cc, pt-eval.h, pt-except.cc,
pt-except.h, pt-exp.cc, pt-exp.h, pt-fcn-handle.cc, pt-fcn-handle.h, pt-id.cc,
pt-id.h, pt-idx.cc, pt-idx.h, pt-jump.h, pt-loop.cc, pt-loop.h, pt-mat.cc,
pt-mat.h, pt-misc.cc, pt-misc.h, pt-pr-code.cc, pt-pr-code.h, pt-select.cc,
pt-select.h, pt-spmd.cc, pt-spmd.h, pt-stmt.cc, pt-stmt.h, pt-tm-const.cc,
pt-tm-const.h, pt-unop.cc, pt-unop.h, pt-vm-eval.cc, pt-walk.cc, pt-walk.h,
pt.cc, pt.h, token.cc, token.h, Range.cc, Range.h, idx-vector.cc, idx-vector.h,
range-fwd.h, CollocWt.cc, CollocWt.h, aepbalance.cc, aepbalance.h, chol.cc,
chol.h, gepbalance.cc, gepbalance.h, gsvd.cc, gsvd.h, hess.cc, hess.h,
lo-mappers.cc, lo-mappers.h, lo-specfun.cc, lo-specfun.h, lu.cc, lu.h,
oct-convn.cc, oct-convn.h, oct-fftw.cc, oct-fftw.h, oct-norm.cc, oct-norm.h,
oct-rand.cc, oct-rand.h, oct-spparms.cc, oct-spparms.h, qr.cc, qr.h, qrp.cc,
qrp.h, randgamma.cc, randgamma.h, randmtzig.cc, randmtzig.h, randpoisson.cc,
randpoisson.h, schur.cc, schur.h, sparse-chol.cc, sparse-chol.h, sparse-lu.cc,
sparse-lu.h, sparse-qr.cc, sparse-qr.h, svd.cc, svd.h, child-list.cc,
child-list.h, dir-ops.cc, dir-ops.h, file-ops.cc, file-ops.h, file-stat.cc,
file-stat.h, lo-sysdep.cc, lo-sysdep.h, lo-sysinfo.cc, lo-sysinfo.h,
mach-info.cc, mach-info.h, oct-env.cc, oct-env.h, oct-group.cc, oct-group.h,
oct-password.cc, oct-password.h, oct-syscalls.cc, oct-syscalls.h, oct-time.cc,
oct-time.h, oct-uname.cc, oct-uname.h, action-container.cc, action-container.h,
base-list.h, cmd-edit.cc, cmd-edit.h, cmd-hist.cc, cmd-hist.h, f77-fcn.h,
file-info.cc, file-info.h, lo-array-errwarn.cc, lo-array-errwarn.h, lo-hash.cc,
lo-hash.h, lo-ieee.h, lo-regexp.cc, lo-regexp.h, lo-utils.cc, lo-utils.h,
oct-base64.cc, oct-base64.h, oct-glob.cc, oct-glob.h, oct-inttypes.h,
oct-mutex.cc, oct-mutex.h, oct-refcount.h, oct-shlib.cc, oct-shlib.h,
oct-sparse.cc, oct-sparse.h, oct-string.h, octave-preserve-stream-state.h,
pathsearch.cc, pathsearch.h, quit.cc, quit.h, unwind-prot.cc, unwind-prot.h,
url-transfer.cc, url-transfer.h : Use new macros to begin/end C++ namespaces.
author | Rik <rik@octave.org> |
---|---|
date | Thu, 01 Dec 2022 14:23:45 -0800 |
parents | 9f4a9dd4a6ee |
children | aac27ad79be6 |
line wrap: on
line source
//////////////////////////////////////////////////////////////////////// // // Copyright (C) 2002-2022 The Octave Project Developers // // See the file COPYRIGHT.md in the top-level directory of this // distribution or <https://octave.org/copyright/>. // // This file is part of Octave. // // Octave is free software: you can redistribute it and/or modify it // under the terms of the GNU General Public License as published by // the Free Software Foundation, either version 3 of the License, or // (at your option) any later version. // // Octave is distributed in the hope that it will be useful, but // WITHOUT ANY WARRANTY; without even the implied warranty of // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the // GNU General Public License for more details. // // You should have received a copy of the GNU General Public License // along with Octave; see the file COPYING. If not, see // <https://www.gnu.org/licenses/>. // //////////////////////////////////////////////////////////////////////// #if defined (HAVE_CONFIG_H) # include "config.h" #endif #include <list> #include <sstream> #include <string> #include <vector> #if defined (HAVE_PCRE2_H) || defined (HAVE_PCRE2_PCRE2_H) # define PCRE2_CODE_UNIT_WIDTH 8 # if defined (HAVE_PCRE2_H) # include <pcre2.h> # elif defined (HAVE_PCRE2_PCRE2_H) # include <pcre2/pcre2.h> # endif #elif defined (HAVE_PCRE_H) || defined (HAVE_PCRE_PCRE_H) # if defined (HAVE_PCRE_H) # include <pcre.h> # elif defined (HAVE_PCRE_PCRE_H) # include <pcre/pcre.h> # endif #endif #include "Matrix.h" #include "base-list.h" #include "lo-error.h" #include "oct-locbuf.h" #include "quit.h" #include "lo-regexp.h" #include "str-vec.h" #include "unistr-wrappers.h" #include "unwind-prot.h" #if defined (HAVE_PCRE2) typedef pcre2_code octave_pcre_code; typedef PCRE2_SIZE OCTAVE_PCRE_SIZE; void (*octave_pcre_code_free) (octave_pcre_code *) = pcre2_code_free; # define OCTAVE_PCRE_CASELESS PCRE2_CASELESS # define OCTAVE_PCRE_DOTALL PCRE2_DOTALL # define OCTAVE_PCRE_MULTILINE PCRE2_MULTILINE # define OCTAVE_PCRE_EXTENDED PCRE2_EXTENDED # define OCTAVE_PCRE_UTF PCRE2_UTF # define OCTAVE_PCRE_INFO_CAPTURECOUNT PCRE2_INFO_CAPTURECOUNT # define OCTAVE_PCRE_INFO_NAMECOUNT PCRE2_INFO_NAMECOUNT # define OCTAVE_PCRE_INFO_NAMEENTRYSIZE PCRE2_INFO_NAMEENTRYSIZE # define OCTAVE_PCRE_INFO_NAMETABLE PCRE2_INFO_NAMETABLE #elif defined (HAVE_PCRE) typedef pcre octave_pcre_code; typedef int OCTAVE_PCRE_SIZE; void (*octave_pcre_code_free) (void *) = pcre_free; # define OCTAVE_PCRE_CASELESS PCRE_CASELESS # define OCTAVE_PCRE_DOTALL PCRE_DOTALL # define OCTAVE_PCRE_MULTILINE PCRE_MULTILINE # define OCTAVE_PCRE_EXTENDED PCRE_EXTENDED # define OCTAVE_PCRE_UTF PCRE_UTF8 # define OCTAVE_PCRE_INFO_CAPTURECOUNT PCRE_INFO_CAPTURECOUNT # define OCTAVE_PCRE_INFO_NAMECOUNT PCRE_INFO_NAMECOUNT # define OCTAVE_PCRE_INFO_NAMEENTRYSIZE PCRE_INFO_NAMEENTRYSIZE # define OCTAVE_PCRE_INFO_NAMETABLE PCRE_INFO_NAMETABLE #else # error "PCRE2 or PCRE library is required to build Octave" #endif static inline int octave_pcre_pattern_info (const octave_pcre_code *code, int what, void *where) { #if defined (HAVE_PCRE2) return pcre2_pattern_info (code, what, where); #else return pcre_fullinfo (code, nullptr, what, where); #endif } OCTAVE_BEGIN_NAMESPACE(octave) // Define the maximum number of retries for a pattern // that possibly results in an infinite recursion. #define PCRE_MATCHLIMIT_MAX 10 // FIXME: should this be configurable? #define MAXLOOKBEHIND 10 static bool lookbehind_warned = false; // FIXME: don't bother collecting and composing return values // the user doesn't want. void regexp::free (void) { octave_pcre_code_free (static_cast<octave_pcre_code *> (m_code)); } void regexp::compile_internal (void) { // If we had a previously compiled pattern, release it. free (); std::size_t max_length = MAXLOOKBEHIND; std::size_t pos = 0; std::size_t new_pos; int inames = 0; std::ostringstream buf; while ((new_pos = m_pattern.find ("(?", pos)) != std::string::npos) { std::size_t tmp_pos; if (m_pattern.size () > new_pos + 2 && m_pattern.at (new_pos + 2) == '<' && ! (m_pattern.size () > new_pos + 3 && (m_pattern.at (new_pos + 3) == '=' || m_pattern.at (new_pos + 3) == '!')) && (tmp_pos = m_pattern.find_first_of ('>', new_pos)) != std::string::npos && m_pattern.find_first_of (')', tmp_pos) != std::string::npos) { // The syntax of named tokens in pcre is "(?P<name>...)" while // we need a syntax "(?<name>...)", so fix that here. Also an // expression like // "(?<first>\w+)\s+(?<last>\w+)|(?<last>\w+),\s+(?<first>\w+)" // should be perfectly legal, while pcre does not allow the same // named token name on both sides of the alternative. Also fix // that here by replacing name tokens by dummy names, and dealing // with the dummy names later. std::string tmp_name = m_pattern.substr (new_pos+3, tmp_pos-new_pos-3); bool found = false; for (int i = 0; i < m_names; i++) { if (m_named_pats(i) == tmp_name) { m_named_idx.resize (dim_vector (inames+1, 1)); m_named_idx(inames) = i; found = true; break; } } if (! found) { m_named_idx.resize (dim_vector (inames+1, 1)); m_named_idx(inames) = m_names; m_named_pats.append (tmp_name); m_names++; } if (new_pos - pos > 0) buf << m_pattern.substr (pos, new_pos-pos); if (inames < 10) buf << "(?P<n00" << inames++; else if (inames < 100) buf << "(?P<n0" << inames++; else buf << "(?P<n" << inames++; pos = tmp_pos; } else if (m_pattern.size () > new_pos + 2 && m_pattern.at (new_pos + 2) == '<') { // Find lookbehind operators of arbitrary length (ie like // "(?<=[a-z]*)") and replace with a maximum length operator // as PCRE can not yet handle arbitrary length lookahead // operators. Use the string length as the maximum length to // avoid issues. int brackets = 1; std::size_t tmp_pos1 = new_pos + 2; std::size_t tmp_pos2 = tmp_pos1; while (tmp_pos1 < m_pattern.length () && brackets > 0) { char ch = m_pattern.at (tmp_pos1); if (ch == '(') brackets++; else if (ch == ')') { if (brackets > 1) tmp_pos2 = tmp_pos1; brackets--; } tmp_pos1++; } if (brackets != 0) { buf << m_pattern.substr (pos, new_pos - pos) << "(?"; pos = new_pos + 2; } else { std::size_t tmp_pos3 = m_pattern.find_first_of ("*+", tmp_pos2); if (tmp_pos3 != std::string::npos && tmp_pos3 < tmp_pos1) { if (! lookbehind_warned) { lookbehind_warned = true; (*current_liboctave_warning_with_id_handler) ("Octave:regexp-lookbehind-limit", "%s: arbitrary length lookbehind patterns are only supported up to length %d", m_who.c_str (), MAXLOOKBEHIND); } buf << m_pattern.substr (pos, new_pos - pos) << '('; std::size_t i; if (m_pattern.at (tmp_pos3) == '*') i = 0; else i = 1; for (; i < max_length + 1; i++) { buf << m_pattern.substr (new_pos, tmp_pos3 - new_pos) << '{' << i << '}'; buf << m_pattern.substr (tmp_pos3 + 1, tmp_pos1 - tmp_pos3 - 1); if (i != max_length) buf << '|'; } buf << ')'; } else buf << m_pattern.substr (pos, tmp_pos1 - pos); pos = tmp_pos1; } } else { buf << m_pattern.substr (pos, new_pos - pos) << "(?"; pos = new_pos + 2; } } buf << m_pattern.substr (pos); // Replace NULLs with escape sequence because conversion function c_str() // will terminate string early at embedded NULLs. std::string buf_str = buf.str (); while ((pos = buf_str.find ('\0')) != std::string::npos) buf_str.replace (pos, 1, "\\000"); int pcre_options = ( (m_options.case_insensitive () ? OCTAVE_PCRE_CASELESS : 0) | (m_options.dotexceptnewline () ? 0 : OCTAVE_PCRE_DOTALL) | (m_options.lineanchors () ? OCTAVE_PCRE_MULTILINE : 0) | (m_options.freespacing () ? OCTAVE_PCRE_EXTENDED : 0) | OCTAVE_PCRE_UTF); #if defined (HAVE_PCRE2) PCRE2_SIZE erroffset; int errnumber; m_code = pcre2_compile (reinterpret_cast<PCRE2_SPTR> (buf_str.c_str ()), PCRE2_ZERO_TERMINATED, pcre_options, &errnumber, &erroffset, nullptr); if (! m_code) { // PCRE docs say: // // If the buffer is too small, the message is truncated (but // still with a trailing zero), and the negative error code // PCRE2_ERROR_NOMEMORY is returned. None of the messages are // very long; a buffer size of 120 code units is ample. // // so we assume that 256 will be large enough to avoid truncated // messages. PCRE2_UCHAR err [256]; pcre2_get_error_message (errnumber, err, sizeof (err)); (*current_liboctave_error_handler) ("%s: %s at position %zu of expression", m_who.c_str (), err, erroffset); } #else const char *err; int erroffset; m_code = pcre_compile (buf_str.c_str (), pcre_options, &err, &erroffset, nullptr); if (! m_code) (*current_liboctave_error_handler) ("%s: %s at position %d of expression", m_who.c_str (), err, erroffset); #endif } regexp::match_data regexp::match (const std::string& buffer) const { // check if input is valid utf-8 const uint8_t *buf_str = reinterpret_cast<const uint8_t *> (buffer.c_str ()); if (octave_u8_check_wrapper (buf_str, buffer.length ())) (*current_liboctave_error_handler) ("%s: the input string is invalid UTF-8", m_who.c_str ()); regexp::match_data retval; std::list<regexp::match_element> lst; int subpatterns; int namecount; int nameentrysize; char *nametable; std::size_t idx = 0; octave_pcre_code *re = static_cast<octave_pcre_code *> (m_code); octave_pcre_pattern_info (re, OCTAVE_PCRE_INFO_CAPTURECOUNT, &subpatterns); octave_pcre_pattern_info (re, OCTAVE_PCRE_INFO_NAMECOUNT, &namecount); octave_pcre_pattern_info (re, OCTAVE_PCRE_INFO_NAMEENTRYSIZE, &nameentrysize); octave_pcre_pattern_info (re, OCTAVE_PCRE_INFO_NAMETABLE, &nametable); #if defined (HAVE_PCRE) OCTAVE_LOCAL_BUFFER (OCTAVE_PCRE_SIZE, ovector, (subpatterns+1)*3); #endif OCTAVE_LOCAL_BUFFER (int, nidx, namecount); for (int i = 0; i < namecount; i++) { // Index of subpattern in first two bytes of name (MSB first). // Extract index. nidx[i] = (static_cast<int> (nametable[i*nameentrysize])) << 8 | static_cast<int> (nametable[i*nameentrysize+1]); } while (true) { octave_quit (); #if defined (HAVE_PCRE2) pcre2_match_data *m_data = pcre2_match_data_create_from_pattern (re, nullptr); unwind_action cleanup_match_data ([=] () { pcre2_match_data_free (m_data); }); int matches = pcre2_match (re, reinterpret_cast<PCRE2_SPTR> (buffer.c_str ()), buffer.length (), idx, PCRE2_NO_UTF_CHECK | (idx ? PCRE2_NOTBOL : 0), m_data, nullptr); if (matches < 0 && matches != PCRE2_ERROR_NOMATCH) (*current_liboctave_error_handler) ("%s: internal error calling pcre2_match; " "error code from pcre2_match is %i", m_who.c_str (), matches); if (matches == PCRE2_ERROR_NOMATCH) break; OCTAVE_PCRE_SIZE *ovector = pcre2_get_ovector_pointer (m_data); #else int matches = pcre_exec (re, nullptr, buffer.c_str (), buffer.length (), idx, PCRE_NO_UTF8_CHECK | (idx ? PCRE_NOTBOL : 0), ovector, (subpatterns+1)*3); if (matches == PCRE_ERROR_MATCHLIMIT) { // Try harder; start with default value for MATCH_LIMIT // and increase it. (*current_liboctave_warning_with_id_handler) ("Octave:regexp-match-limit", "your pattern caused PCRE to hit its MATCH_LIMIT; trying harder now, but this will be slow"); pcre_extra pe; pcre_config (PCRE_CONFIG_MATCH_LIMIT, static_cast<void *> (&pe.match_limit)); pe.flags = PCRE_EXTRA_MATCH_LIMIT; int i = 0; while (matches == PCRE_ERROR_MATCHLIMIT && i++ < PCRE_MATCHLIMIT_MAX) { octave_quit (); pe.match_limit *= 10; matches = pcre_exec (re, &pe, buffer.c_str (), buffer.length (), idx, PCRE_NO_UTF8_CHECK | (idx ? PCRE_NOTBOL : 0), ovector, (subpatterns+1)*3); } } if (matches < 0 && matches != PCRE_ERROR_NOMATCH) (*current_liboctave_error_handler) ("%s: internal error calling pcre_exec; " "error code from pcre_exec is %i", m_who.c_str (), matches); if (matches == PCRE_ERROR_NOMATCH) break; #endif if (ovector[0] >= ovector[1] && ! m_options.emptymatch ()) { // Zero length match. Skip to next char. idx = ovector[0] + 1; if (idx < buffer.length ()) continue; else break; } else { int pos_match = 0; Matrix token_extents (matches-1, 2); for (int i = 1; i < matches; i++) { #if defined (HAVE_PCRE2) if (ovector[2*i] != PCRE2_SIZE_MAX #else if (ovector[2*i] >= 0 #endif && ovector[2*i+1] > 0 && (i == 1 || ovector[2*i] != ovector[2*i-2] || ovector[2*i-1] != ovector[2*i+1])) { token_extents(pos_match, 0) = double (ovector[2*i]+1); token_extents(pos_match++, 1) = double (ovector[2*i+1]); } } token_extents.resize (pos_match, 2); OCTAVE_PCRE_SIZE start = ovector[0] + 1; OCTAVE_PCRE_SIZE end = ovector[1]; #if defined (HAVE_PCRE2) // Must use explicit length constructor as match can contain '\0'. std::string match_string = std::string (buffer.c_str() + start - 1, end - start + 1); #else const char **listptr; int status = pcre_get_substring_list (buffer.c_str (), ovector, matches, &listptr); if (status == PCRE_ERROR_NOMEMORY) (*current_liboctave_error_handler) ("%s: cannot allocate memory in pcre_get_substring_list", m_who.c_str ()); // Must use explicit length constructor as match can contain '\0'. std::string match_string = std::string (*listptr, end - start + 1); #endif string_vector tokens (pos_match); string_vector named_tokens (m_names); int pos_offset = 0; pos_match = 0; for (int i = 1; i < matches; i++) { #if defined (HAVE_PCRE2) if (ovector[2*i] != PCRE2_SIZE_MAX #else if (ovector[2*i] >= 0 #endif && ovector[2*i+1] > 0) { if (i == 1 || ovector[2*i] != ovector[2*i-2] || ovector[2*i-1] != ovector[2*i+1]) { if (namecount > 0) { // FIXME: Should probably do this with a map() // rather than a linear search. However, // the number of captured, named expressions // is usually pretty small (< 4) for (int j = 0; j < namecount; j++) { if (nidx[j] == i) { std::size_t len = ovector[2*i+1] - ovector[2*i]; named_tokens(m_named_idx(j)) #if defined (HAVE_PCRE2) = std::string (buffer.c_str () + ovector[2*i], len); #else = std::string (*(listptr+i-pos_offset), len); #endif break; } } } std::size_t len = ovector[2*i+1] - ovector[2*i]; #if defined (HAVE_PCRE2) tokens(pos_match++) = std::string (buffer.c_str() + ovector[2*i], len); #else tokens(pos_match++) = std::string (*(listptr+i), len); #endif } else pos_offset++; } } #if ! defined (HAVE_PCRE2) pcre_free_substring_list (listptr); #endif // FIXME: MATCH_ELEMENT uses double values for these, // presumably because that is what the Octave interpreter // uses. Should we check that the values don't exceed // flintmax here? It seems unlikely that it would happen, // but... double dstart = static_cast<double> (start); double dend = static_cast<double> (end); regexp::match_element new_elem (named_tokens, tokens, match_string, token_extents, dstart, dend); lst.push_back (new_elem); if (ovector[1] <= ovector[0]) { // Zero length match. Skip to next char. idx = ovector[0] + 1; if (idx <= buffer.length ()) continue; } else idx = ovector[1]; if (m_options.once () || idx >= buffer.length ()) break; } } retval = regexp::match_data (lst, m_named_pats); return retval; } bool regexp::is_match (const std::string& buffer) const { regexp::match_data rx_lst = match (buffer); return rx_lst.size () > 0; } Array<bool> regexp::is_match (const string_vector& buffer) const { octave_idx_type len = buffer.numel (); Array<bool> retval (dim_vector (len, 1)); for (octave_idx_type i = 0; i < buffer.numel (); i++) retval(i) = is_match (buffer(i)); return retval; } // Declare rep_token_t used in processing replacement string struct rep_token_t { std::size_t pos; int num; }; std::string regexp::replace (const std::string& buffer, const std::string& replacement) const { std::string retval; const regexp::match_data rx_lst = match (buffer); std::size_t num_matches = rx_lst.size (); if (num_matches == 0) { retval = buffer; return retval; } // Identify replacement tokens; build a vector of group numbers in // the replacement string so that we can quickly calculate the size // of the replacement. // FIXME: All code assumes that only 10 tokens ($0-$9) exist. // $11 represents $1 followed by the character '1' rather than // the eleventh capture buffer. std::string repstr = replacement; std::vector<rep_token_t> tokens; tokens.reserve (5); // Reserve memory for 5 pattern replacements for (std::size_t i=0; i < repstr.size (); i++) { if (repstr[i] == '\\') { if (i < repstr.size () - 1 && repstr[i+1] == '$') { repstr.erase (i, 1); // erase backslash i++; // skip over '$' continue; } if (i < repstr.size () - 1 && repstr[i+1] == '\\') { repstr.erase (i, 1); // erase 1st backslash continue; } } else if (repstr[i] == '$') { if (i < repstr.size () - 1 && isdigit (repstr[i+1])) { rep_token_t tmp_token; tmp_token.pos = i; tmp_token.num = repstr[i+1]-'0'; tokens.push_back (tmp_token); } } } std::string rep; int num_tokens = tokens.size (); if (num_tokens > 0) { // Determine replacement length const std::size_t replen = repstr.size () - 2*num_tokens; int delta = 0; auto p = rx_lst.begin (); for (std::size_t i = 0; i < num_matches; i++) { octave_quit (); double start = p->start (); double end = p->end (); const Matrix pairs (p->token_extents ()); std::size_t pairlen = 0; for (int j = 0; j < num_tokens; j++) { if (tokens[j].num == 0) pairlen += static_cast<std::size_t> (end - start + 1); else if (tokens[j].num <= pairs.rows ()) pairlen += static_cast<std::size_t> (pairs(tokens[j].num-1,1) - pairs(tokens[j].num-1,0) + 1); } delta += (static_cast<int> (replen + pairlen) - static_cast<int> (end - start + 1)); p++; } // Build replacement string rep.reserve (buffer.size () + delta); std::size_t from = 0; p = rx_lst.begin (); for (std::size_t i = 0; i < num_matches; i++) { octave_quit (); double start = p->start (); double end = p->end (); const Matrix pairs (p->token_extents ()); rep.append (&buffer[from], static_cast<std::size_t> (start - 1 - from)); from = static_cast<std::size_t> (end); std::size_t cur_pos = 0; for (int j = 0; j < num_tokens; j++) { rep.append (&repstr[cur_pos], (tokens[j].pos) - cur_pos); cur_pos = tokens[j].pos+2; int k = tokens[j].num; if (k == 0) { // replace with entire match rep.append (&buffer[static_cast<std::size_t> (end - 1)], static_cast<std::size_t> (end - start + 1)); } else if (k <= pairs.rows ()) { // replace with group capture rep.append (&buffer[static_cast<std::size_t> (pairs(k-1,0)-1)], static_cast<std::size_t> (pairs(k-1,1) - pairs(k-1,0) + 1)); } else { // replace with nothing } } if (cur_pos < repstr.size ()) rep.append (&repstr[cur_pos], repstr.size () - cur_pos); p++; } rep.append (&buffer[from], buffer.size () - from); } else { // Determine repstr length const std::size_t replen = repstr.size (); int delta = 0; auto p = rx_lst.begin (); for (std::size_t i = 0; i < num_matches; i++) { octave_quit (); delta += static_cast<int> (replen) - static_cast<int> (p->end () - p->start () + 1); p++; } // Build replacement string rep.reserve (buffer.size () + delta); std::size_t from = 0; p = rx_lst.begin (); for (std::size_t i = 0; i < num_matches; i++) { octave_quit (); rep.append (&buffer[from], static_cast<std::size_t> (p->start () - 1 - from)); from = static_cast<std::size_t> (p->end ()); rep.append (repstr); p++; } rep.append (&buffer[from], buffer.size () - from); } retval = rep; return retval; } OCTAVE_END_NAMESPACE(octave)