Now Adv. F&R correctly interprets escape sequences within regexp insets, e.g., \\ or \(, for finding characters with a special meaning in regular expression context (\^()[]).

Added corresponding regression tests findadv-re-01 and -02 for finding these special characters.
Unfortunately, braces {} are a little bit special and don't work yet. Added findadv-re-03 regression test to not forget about this case.
This goes along the line of fixing issues reported by Andrew Parsloe and summarised as #7638.
Last note: Still there will be cases when the Advanced F&R won't work, especially if the text to be found and/or the search pattern is preceded by backslashes or others.


git-svn-id: svn://svn.lyx.org/lyx/lyx-devel/trunk@39135 a592a061-630c-0410-9148-cb99ea01b6c8
This commit is contained in:
Tommaso Cucinotta 2011-06-19 16:41:23 +00:00
parent 008815a972
commit a48baee56f
5 changed files with 224 additions and 45 deletions

View File

@ -0,0 +1,35 @@
# Finding special latex/regexp chars: (
#
Lang it_IT.utf8
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \\()[]{}%^#@?*.\C\[Home]
KK: \Cs
KK: \CF
KK: \Axregexp-mode\[Return]\\(\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 1\n with len: 1' lyx-log.txt
#
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
# Uncheck ignore format
KK: \Az\Ag\Ae
KK: \Axregexp-mode\[Return]\\(\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 1\n with len: 1' lyx-log.txt
# Finding special latex/regexp chars: [
#
Lang it_IT.utf8
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
KK: \Axregexp-mode\[Return]\\[\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 3\n with len: 1' lyx-log.txt
#
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
# Uncheck ignore format
KK: \Az\Ag\Ae
KK: \Axregexp-mode\[Return]\\[\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 3\n with len: 1' lyx-log.txt

View File

@ -0,0 +1,52 @@
# Finding special latex/regexp chars: \
#
Lang it_IT.utf8
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \\()[]{}%^#@?*.\C\[Home]
KK: \Cs
KK: \CF
KK: \Axregexp-mode\[Return]\\\\\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 0\n with len: 1' lyx-log.txt
#
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
# Uncheck ignore format
KK: \Az\Ag\Ae
KK: \Axregexp-mode\[Return]\\\\\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 0\n with len: 1' lyx-log.txt
# Finding special latex/regexp chars: %
#
Lang it_IT.utf8
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
KK: \Axregexp-mode\[Return]%\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 7\n with len: 1' lyx-log.txt
#
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
# Uncheck ignore format
KK: \Az\Ag\Ae
KK: \Axregexp-mode\[Return]%\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 7\n with len: 1' lyx-log.txt
# Finding special latex/regexp chars: ^
#
Lang it_IT.utf8
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
KK: \Axregexp-mode\[Return]\\^\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 8\n with len: 1' lyx-log.txt
#
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
# Uncheck ignore format
KK: \Az\Ag\Ae
KK: \Axregexp-mode\[Return]\\^\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 8\n with len: 1' lyx-log.txt

View File

@ -0,0 +1,35 @@
# Finding special latex/regexp chars: {
#
Lang it_IT.utf8
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \\()[]{}%^#@?*.\C\[Home]
KK: \Cs
KK: \CF
KK: \Axregexp-mode\[Return]\\{\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 5\n with len: 1' lyx-log.txt
#
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
# Uncheck ignore format
KK: \Az\Ag\Ae
KK: \Axregexp-mode\[Return]\\{\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 5\n with len: 1' lyx-log.txt
# Finding special latex/regexp chars: }
#
Lang it_IT.utf8
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
KK: \Axregexp-mode\[Return]\\}\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 5\n with len: 1' lyx-log.txt
#
TestBegin test.lyx -dbg find > lyx-log.txt 2>&1
KK: \CF
# Uncheck ignore format
KK: \Az\Ag\Ae
KK: \Axregexp-mode\[Return]\\}\[Return]
TestEnd
Assert pcregrep -M 'Putting selection at .*idx: 0 par: 0 pos: 5\n with len: 1' lyx-log.txt

View File

@ -490,12 +490,11 @@ namespace {
typedef vector<pair<string, string> > Escapes; typedef vector<pair<string, string> > Escapes;
/// A map of symbols and their escaped equivalent needed within a regex. /// A map of symbols and their escaped equivalent needed within a regex.
/// @note Beware of order
Escapes const & get_regexp_escapes() Escapes const & get_regexp_escapes()
{ {
static Escapes escape_map; static Escapes escape_map;
if (escape_map.empty()) { if (escape_map.empty()) {
escape_map.push_back(pair<string, string>("\\", "\\\\"));
escape_map.push_back(pair<string, string>("^", "\\^"));
escape_map.push_back(pair<string, string>("$", "\\$")); escape_map.push_back(pair<string, string>("$", "\\$"));
escape_map.push_back(pair<string, string>("{", "\\{")); escape_map.push_back(pair<string, string>("{", "\\{"));
escape_map.push_back(pair<string, string>("}", "\\}")); escape_map.push_back(pair<string, string>("}", "\\}"));
@ -506,6 +505,9 @@ Escapes const & get_regexp_escapes()
escape_map.push_back(pair<string, string>("+", "\\+")); escape_map.push_back(pair<string, string>("+", "\\+"));
escape_map.push_back(pair<string, string>("*", "\\*")); escape_map.push_back(pair<string, string>("*", "\\*"));
escape_map.push_back(pair<string, string>(".", "\\.")); escape_map.push_back(pair<string, string>(".", "\\."));
escape_map.push_back(pair<string, string>("\\", "(?:\\\\|\\\\backslash)"));
escape_map.push_back(pair<string, string>("~", "(?:\\\\textasciitilde|\\\\sim)"));
escape_map.push_back(pair<string, string>("^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\mathcircumflex"));
} }
return escape_map; return escape_map;
} }
@ -514,14 +516,30 @@ Escapes const & get_regexp_escapes()
Escapes const & get_lyx_unescapes() { Escapes const & get_lyx_unescapes() {
static Escapes escape_map; static Escapes escape_map;
if (escape_map.empty()) { if (escape_map.empty()) {
escape_map.push_back(pair<string, string>("{*}", "*")); escape_map.push_back(pair<string, string>("\\%", "%"));
escape_map.push_back(pair<string, string>("{[}", "[")); escape_map.push_back(pair<string, string>("\\{", "{"));
escape_map.push_back(pair<string, string>("\\$", "$")); escape_map.push_back(pair<string, string>("\\}", "}"));
escape_map.push_back(pair<string, string>("\\backslash{}", "\\")); escape_map.push_back(pair<string, string>("\\mathcircumflex ", "^"));
escape_map.push_back(pair<string, string>("\\mathcircumflex", "^"));
escape_map.push_back(pair<string, string>("\\backslash ", "\\")); escape_map.push_back(pair<string, string>("\\backslash ", "\\"));
escape_map.push_back(pair<string, string>("\\backslash", "\\")); escape_map.push_back(pair<string, string>("\\backslash", "\\"));
escape_map.push_back(pair<string, string>("\\sim ", "~")); escape_map.push_back(pair<string, string>("\\sim ", "~"));
escape_map.push_back(pair<string, string>("\\^", "^")); escape_map.push_back(pair<string, string>("\\sim", "~"));
}
return escape_map;
}
/// A map of escapes turning a regexp matching text to one matching latex.
Escapes const & get_regexp_latex_escapes() {
static Escapes escape_map;
if (escape_map.empty()) {
escape_map.push_back(pair<string, string>("\\\\", "(?:\\\\\\\\|\\\\backslash|\\\\textbackslash\\{\\})"));
escape_map.push_back(pair<string, string>("(<?!\\\\\\\\textbackslash)\\{", "\\\\\\{"));
escape_map.push_back(pair<string, string>("(<?!\\\\\\\\textbackslash\\\\\\{)\\}", "\\\\\\}"));
escape_map.push_back(pair<string, string>("\\[", "\\{\\[\\}"));
escape_map.push_back(pair<string, string>("\\]", "\\{\\]\\}"));
escape_map.push_back(pair<string, string>("\\^", "(?:\\^|\\\\textasciicircum\\{\\}|\\\\mathcircumflex)"));
escape_map.push_back(pair<string, string>("%", "\\\\\\%"));
} }
return escape_map; return escape_map;
} }
@ -538,7 +556,7 @@ string apply_escapes(string s, Escapes const & escape_map)
unsigned int pos = 0; unsigned int pos = 0;
while (pos < s.length() && (pos = s.find(it->first, pos)) < s.length()) { while (pos < s.length() && (pos = s.find(it->first, pos)) < s.length()) {
s.replace(pos, it->first.length(), it->second); s.replace(pos, it->first.length(), it->second);
// LYXERR(Debug::FIND, "After escape: " << s); LYXERR(Debug::FIND, "After escape: " << s);
pos += it->second.length(); pos += it->second.length();
// LYXERR(Debug::FIND, "pos: " << pos); // LYXERR(Debug::FIND, "pos: " << pos);
} }
@ -550,7 +568,7 @@ string apply_escapes(string s, Escapes const & escape_map)
/** Return the position of the closing brace matching the open one at s[pos], /** Return the position of the closing brace matching the open one at s[pos],
** or s.size() if not found. ** or s.size() if not found.
**/ **/
size_t find_matching_brace(string const & s, size_t pos) static size_t find_matching_brace(string const & s, size_t pos)
{ {
LASSERT(s[pos] == '{', /* */); LASSERT(s[pos] == '{', /* */);
int open_braces = 1; int open_braces = 1;
@ -570,7 +588,8 @@ size_t find_matching_brace(string const & s, size_t pos)
/// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string), /// Within \regexp{} apply get_lyx_unescapes() only (i.e., preserve regexp semantics of the string),
/// while outside apply get_lyx_unescapes()+get_regexp_escapes(). /// while outside apply get_lyx_unescapes()+get_regexp_escapes().
string escape_for_regex(string s) /// If match_latex is true, then apply regexp_latex_escapes() to \regexp{} contents as well.
string escape_for_regex(string s, bool match_latex)
{ {
size_t pos = 0; size_t pos = 0;
while (pos < s.size()) { while (pos < s.size()) {
@ -591,6 +610,8 @@ string escape_for_regex(string s)
size_t end_pos = s.find("}}}", new_pos + 10); // find_matching_brace(s, new_pos + 7); size_t end_pos = s.find("}}}", new_pos + 10); // find_matching_brace(s, new_pos + 7);
LYXERR(Debug::FIND, "end_pos: " << end_pos); LYXERR(Debug::FIND, "end_pos: " << end_pos);
t = apply_escapes(s.substr(new_pos + 10, end_pos - (new_pos + 10)), get_lyx_unescapes()); t = apply_escapes(s.substr(new_pos + 10, end_pos - (new_pos + 10)), get_lyx_unescapes());
if (match_latex)
t = apply_escapes(t, get_regexp_latex_escapes());
LYXERR(Debug::FIND, "t : " << t); LYXERR(Debug::FIND, "t : " << t);
if (end_pos == s.size()) { if (end_pos == s.size()) {
s.replace(new_pos, end_pos - new_pos, t); s.replace(new_pos, end_pos - new_pos, t);
@ -711,6 +732,10 @@ private:
lyx::regex regexp; lyx::regex regexp;
// same as regexp, but prefixed with a ".*" // same as regexp, but prefixed with a ".*"
lyx::regex regexp2; lyx::regex regexp2;
// leading format material as string
string lead_as_string;
// par_as_string after removal of lead_as_string
string par_as_string_nolead;
// unmatched open braces in the search string/regexp // unmatched open braces in the search string/regexp
int open_braces; int open_braces;
// number of (.*?) subexpressions added at end of search regexp for closing // number of (.*?) subexpressions added at end of search regexp for closing
@ -769,7 +794,7 @@ static docstring stringifySearchBuffer(Buffer & buffer, FindAndReplaceOptions co
static size_t identifyLeading(string const & s) { static size_t identifyLeading(string const & s) {
string t = s; string t = s;
// @TODO Support \item[text] // @TODO Support \item[text]
while (regex_replace(t, t, "\\\\(emph|textbf|subsubsection|subsection|section|subparagraph|paragraph|part)\\*?\\{", "") while (regex_replace(t, t, "^\\\\(emph|textbf|subsubsection|subsection|section|subparagraph|paragraph|part)\\*?\\{", "")
|| regex_replace(t, t, "^\\$", "") || regex_replace(t, t, "^\\$", "")
|| regex_replace(t, t, "^\\\\\\[ ", "") || regex_replace(t, t, "^\\\\\\[ ", "")
|| regex_replace(t, t, "^\\\\item ", "") || regex_replace(t, t, "^\\\\item ", "")
@ -779,6 +804,27 @@ static size_t identifyLeading(string const & s) {
} }
// Remove trailing closure of math, macros and environments, so to catch parts of them.
static int identifyClosing(string & t) {
int open_braces = 0;
do {
LYXERR(Debug::FIND, "identifyClosing(): t now is '" << t << "'");
if (regex_replace(t, t, "(.*[^\\\\])\\$\\'", "$1"))
continue;
if (regex_replace(t, t, "(.*[^\\\\]) \\\\\\]\\'", "$1"))
continue;
if (regex_replace(t, t, "(.*[^\\\\]) \\\\end\\{[a-zA-Z_]*\\*?\\}\\'", "$1"))
continue;
if (regex_replace(t, t, "(.*[^\\\\])\\}\\'", "$1")) {
++open_braces;
continue;
}
break;
} while (true);
return open_braces;
}
MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt) MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const & opt)
: p_buf(&buf), p_first_buf(&buf), opt(opt) : p_buf(&buf), p_first_buf(&buf), opt(opt)
{ {
@ -787,37 +833,30 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const &
open_braces = 0; open_braces = 0;
close_wildcards = 0; close_wildcards = 0;
use_regexp = !opt.ignoreformat || par_as_string.find("\\regexp") != std::string::npos; use_regexp = par_as_string.find("\\regexp") != std::string::npos;
size_t lead_size = 0;
if (!opt.ignoreformat) {
lead_size = identifyLeading(par_as_string);
lead_as_string = par_as_string.substr(0, lead_size);
par_as_string_nolead = par_as_string.substr(lead_size, par_as_string.size() - lead_size);
}
if (!use_regexp) { if (!use_regexp) {
// Remove trailing closure of math, macros and environments, so to catch parts of them. open_braces = identifyClosing(par_as_string);
do { identifyClosing(par_as_string_nolead);
LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\])\\$\\'", "$1"))
continue;
// @todo need to account for open square braces as well ?
if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) \\\\\\]\\'", "$1"))
continue;
if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\]) \\\\end\\{[a-zA-Z_]*\\*?\\}\\'", "$1"))
continue;
if (regex_replace(par_as_string, par_as_string, "(.*[^\\\\])\\}\\'", "$1")) {
++open_braces;
continue;
}
break;
} while (true);
LYXERR(Debug::FIND, "Open braces: " << open_braces); LYXERR(Debug::FIND, "Open braces: " << open_braces);
LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'"); LYXERR(Debug::FIND, "Built MatchStringAdv object: par_as_string = '" << par_as_string << "'");
} else { } else {
size_t lead_size = identifyLeading(par_as_string);
string lead_as_regexp; string lead_as_regexp;
if (lead_size > 0) { if (lead_size > 0) {
lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size)); // @todo No need to search for \regexp{} insets in leading material
par_as_string = par_as_string.substr(lead_size, par_as_string.size() - lead_size); lead_as_regexp = escape_for_regex(par_as_string.substr(0, lead_size), !opt.ignoreformat);
par_as_string = par_as_string_nolead;
LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'"); LYXERR(Debug::FIND, "lead_as_regexp is '" << lead_as_regexp << "'");
LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
} }
par_as_string = escape_for_regex(par_as_string); par_as_string = escape_for_regex(par_as_string, !opt.ignoreformat);
// Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them. // Insert (.*?) before trailing closure of math, macros and environments, so to catch parts of them.
LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'"); LYXERR(Debug::FIND, "par_as_string now is '" << par_as_string << "'");
if ( if (
@ -839,12 +878,12 @@ MatchStringAdv::MatchStringAdv(lyx::Buffer & buf, FindAndReplaceOptions const &
LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string); LYXERR(Debug::FIND, "Replaced text (to be used as regex): " << par_as_string);
// If entered regexp must match at begin of searched string buffer // If entered regexp must match at begin of searched string buffer
string regexp_str = string("\\`") + lead_as_regexp + par_as_string; string regexp_str = string("\\`") + lead_as_regexp + par_as_string;
LYXERR(Debug::FIND, "Setting regexp to : " << regexp_str << endl); LYXERR(Debug::FIND, "Setting regexp to : '" << regexp_str << "'");
regexp = lyx::regex(regexp_str); regexp = lyx::regex(regexp_str);
// If entered regexp may match wherever in searched string buffer // If entered regexp may match wherever in searched string buffer
string regexp2_str = string("\\`.*") + lead_as_regexp + ".*" + par_as_string; string regexp2_str = string("\\`.*") + lead_as_regexp + ".*" + par_as_string;
LYXERR(Debug::FIND, "Setting regexp2 to: " << regexp2_str << endl); LYXERR(Debug::FIND, "Setting regexp2 to: '" << regexp2_str << "'");
regexp2 = lyx::regex(regexp2_str); regexp2 = lyx::regex(regexp2_str);
} }
} }
@ -858,17 +897,13 @@ int MatchStringAdv::findAux(DocIterator const & cur, int len, bool at_begin) con
LYXERR(Debug::FIND, "After normalization: '" << str << "'"); LYXERR(Debug::FIND, "After normalization: '" << str << "'");
if (! use_regexp) { if (! use_regexp) {
LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='" << par_as_string << "', str='" << str << "'"); LYXERR(Debug::FIND, "Searching in normal mode: par_as_string='" << par_as_string << "', str='" << str << "'");
LYXERR(Debug::FIND, "Searching in normal mode: lead_as_string='" << lead_as_string << "', par_as_string_nolead='" << par_as_string_nolead << "'");
if (at_begin) { if (at_begin) {
LYXERR(Debug::FIND, "size=" << par_as_string.size() << ", substr='" << str.substr(0, par_as_string.size()) << "'"); LYXERR(Debug::FIND, "size=" << par_as_string.size() << ", substr='" << str.substr(0, par_as_string.size()) << "'");
if (str.substr(0, par_as_string.size()) == par_as_string) if (str.substr(0, par_as_string.size()) == par_as_string)
return par_as_string.size(); return par_as_string.size();
} else { } else {
string t = par_as_string; size_t pos = str.find(par_as_string_nolead);
while (regex_replace(t, t, "\\\\(emph|textbf|subsubsection|subsection|section|subparagraph|paragraph|part)\\{", "")
|| regex_replace(t, t, "^\\$", "")
|| regex_replace(t, t, "^\\\\\\[ ", ""))
LYXERR(Debug::FIND, " after removing leading $, \\[ , \\emph{, \\textbf{, etc.: " << t);
size_t pos = str.find(t);
if (pos != string::npos) if (pos != string::npos)
return par_as_string.size(); return par_as_string.size();
} }

View File

@ -1724,7 +1724,29 @@ bool InsetMathNest::interpretChar(Cursor & cur, char_type const c)
} }
// These should be treated differently when not in text mode: // These should be treated differently when not in text mode:
if (currentMode() != InsetMath::TEXT_MODE) { if (cur.inRegexped()) {
switch (c) {
case '\\':
cur.niceInsert(createInsetMath("backslash", buf));
break;
case '^':
cur.niceInsert(createInsetMath("mathcircumflex", buf));
break;
case '{':
case '}':
case '#':
case '%':
case '_':
cur.niceInsert(createInsetMath(docstring(1, c), buf));
break;
case '~':
cur.niceInsert(createInsetMath("sim", buf));
break;
default:
cur.insert(c);
}
return true;
} else if (currentMode() != InsetMath::TEXT_MODE) {
if (c == '_') { if (c == '_') {
script(cur, false, save_selection); script(cur, false, save_selection);
return true; return true;