Commit Graph

69 Commits

Author SHA1 Message Date
Allan Odgaard
e81286b0cd Remove unused include 2014-05-16 22:11:28 +07:00
Allan Odgaard
af3f4aaa33 Check result of most CF/CT ‘Create’ functions 2014-05-15 13:17:30 +07:00
Allan Odgaard
c272afaff2 Cleanup/harmonize whitespace
Leading indent should consist only of tabs, beyond that, only spaces should be used.
2014-04-25 16:55:31 +07:00
Allan Odgaard
9e4a5ccd82 Add instance counter for node_t (used by glob parser)
This will alert us of any leaks (if instance count is non-zero during quit).
2014-04-19 17:01:31 +07:00
Allan Odgaard
91b216bf24 Use glob parser for brace expansion 2014-04-19 17:01:31 +07:00
Allan Odgaard
d178527924 Do not make expansion order assumptions about consecutive braces 2014-04-19 17:01:31 +07:00
Allan Odgaard
7d1f7941b9 Include exclude operator in glob syntax comment/documentation 2014-04-19 17:01:31 +07:00
Allan Odgaard
78e76f55d8 Improve handling of empty braces in globs
We treat empty braces as literal characters though do balance them.
2014-04-18 20:30:07 +07:00
Allan Odgaard
9524439aef Handle bang in globs (negation) identically to tilde
This simplifies the code (slightly) and allows the bang to appear in other positions than leading.
2014-04-18 06:40:48 +07:00
Allan Odgaard
45481d2e4d Add tilde (~) as a glob “exclude” operator
For example the following glob will match C++/Objective-C++ source/header files, except those in the ‘vendor’ directory: *.{cc,mm,h}~vendor/**

It is possible to do multiple exclusions, e.g.: src/**~*.log~*.db

It is also possible to start with the exclude operator to negate the pattern (identical to the ! operator).

The exclude operator is not supported inside brace expansion, e.g.: {*.o,*~.nib} will treat it as a literal tilde. Furthermore, placing the tilde last will also treat it as a literal tilde, e.g.: *.txt~

In the exclude pattern, single/double asterisk (*/**) will match hidden files.
2014-04-18 06:40:48 +07:00
Allan Odgaard
01722b9a6e Create custom parser for glob patterns
Previously we parsed them using regular expressions, but future improvements is hard using that approach.
2014-04-18 06:39:55 +07:00
Allan Odgaard
0b2932b3de Add test for empty glob (should not match anything) 2014-04-18 06:39:28 +07:00
Allan Odgaard
39b94e6ac3 Harmonize whitespace and add trailing newline 2014-04-14 14:26:52 +07:00
Allan Odgaard
96de950ed1 Return initializer list instead of using std::make_pair (C++11) 2014-04-06 19:36:09 +07:00
Allan Odgaard
6185cc17ab Do not inherit numbered variables for potential regexp captures
If the format string itself does a replacement with potential captures $1-$n then we ensure that $1-$n are not inherited from the environment (which could be a parent regexp match).
2014-04-05 14:13:39 +07:00
Allan Odgaard
69c83e56df The ‘const’ keyword should be to the right of its type
The standard doesn’t care which side the keyword is placed on, but placing it on the right makes it easier to read types.

E.g. reading “int const* const” from right to left we get “const pointer to const integer”.
2014-03-31 08:27:19 +07:00
Allan Odgaard
db13d4c1f9 Use std::string::back() instead of operator[] with size()-1 2014-03-31 08:27:18 +07:00
Allan Odgaard
c7ae7d1337 Add test for snippets with overlapping fields
See 4e1cdb4c9e for details.
2014-03-07 00:10:35 +07:00
Allan Odgaard
c2397484b8 Use C++11 for loop
Majority of the edits done using the following ruby script:

    def update_loops(src)
      dst, cnt = '', 0

      block_indent, variable = nil, nil
      src.each_line do |line|
        if block_indent
          if line =~ /^#{block_indent}([{}\t])|^\t*$/
            block_indent = nil if $1 == '}'
            line = line.gsub(%r{ ([^a-z>]) \(\*#{variable}\) | \*#{variable}\b | \b#{variable}(->) }x) do
              $1.to_s + variable + ($2 == "->" ? "." : "")
            end
          else
            block_indent = nil
          end
        elsif line =~ /^(\t*)c?iterate\((\w+), (?!diacritics::make_range)(.*\))$/
          block_indent, variable = $1, $2
          line = "#$1for(auto const& #$2 : #$3\n"
          cnt += 1
        end
        dst << line
      end
      return dst, cnt
    end

    paths.each do |path|
      src = IO.read(path)

      cnt = 1
      while cnt != 0
        src, cnt = update_loops(src)
        STDERR << "#{path}: #{cnt}\n"
      end

      File.open(path, "w") { |io| io << src }
    end
2014-03-03 10:34:13 +07:00
Allan Odgaard
1781f4a605 Use proper iterate macro for temporary 2014-03-03 00:20:02 +07:00
Allan Odgaard
d843225d81 Lines increasing and decreasing indent no longer go below zero
Normally a line matched by both patterns will be decreased by one unit compared to the non-ignored lines immediately above and below. If however the line above was not indented, the line would in theory be decreased to -1 and thus just go back to zero indent for the line following it.
2013-11-08 00:38:53 +01:00
Allan Odgaard
558d7ee8b9 Don’t default placeholders with no default text as mirrors 2013-10-04 16:51:27 +02:00
Allan Odgaard
c5558d0574 Remove a few std::shared_ptr typedefs 2013-10-04 16:51:26 +02:00
Allan Odgaard
1c308c810d Use map::emplace instead of inserting std::pair (C++11) 2013-09-05 20:59:11 +02:00
Allan Odgaard
e4e80a946c Use std::make_shared 2013-09-03 12:27:20 +02:00
Allan Odgaard
27c054484e Allow “stop chars” when parsing a format string
This is useful when we have a format that embed format strings and we wish to use the format string parsers (to avoid an extra layer of escapes).
2013-08-28 15:27:42 +02:00
Allan Odgaard
e1230653f8 Support hex characters in format strings 2013-08-27 15:30:09 +02:00
Allan Odgaard
64663dd17f Allow creating non-capturing regexps 2013-08-25 21:13:06 +02:00
Allan Odgaard
36c5ff4288 Support ${var} in regexp part of format string transformations 2013-08-24 20:28:16 +02:00
Allan Odgaard
1e27f6bf47 Add ONIG_OPTION_NOTGPOS 2013-08-18 17:29:30 +02:00
Allan Odgaard
7350192d51 DFA nodes themselves should not dispose children
A DFA node may be disposed as part of DFA construction (because the node is redundant) and here it would have “next” pointers setup, which it should not follow in its destructor, since those nodes could still be used.

Fixes #1086
2013-08-14 10:41:30 +02:00
Allan Odgaard
5512d65ecf Remove some commented debug output blocks 2013-08-12 22:57:57 +02:00
Allan Odgaard
f7b9694e9f Fix linear destruction of DFA
This is a graph (not a tree) so we have to watch for cycles.
2013-08-12 22:57:57 +02:00
Allan Odgaard
a74bfcc206 Remove some uses of foreach() 2013-08-12 19:32:23 +02:00
Allan Odgaard
f008c49dab Dispose tree structure bottom-up to limit stack usage
With really long search strings, the recursive disposal of nodes would blow the stack
2013-08-10 13:14:55 +02:00
Allan Odgaard
267471d7bb Account for multi-byte characters when moving search position
Fixes #1070
2013-07-31 20:24:16 +02:00
Allan Odgaard
1467a0b44b Update testing system for regexp framework 2013-07-26 13:53:58 +02:00
Allan Odgaard
27eee5ed59 Remove oniguruma 5.9.2 files 2013-07-21 13:25:29 +02:00
Allan Odgaard
e0c7a3df57 Switch regexp library: oniguruma → onigmo 2013-07-21 13:25:29 +02:00
Allan Odgaard
81d065ab8e Use oniguruma’s default character type (OnigUChar)
This introduces a bunch of typecasts but means that less patching is required when updating the library.
2013-07-21 13:25:28 +02:00
Allan Odgaard
8987f3e483 Remove unused function: pattern_t::indices_for_group 2013-07-21 13:25:28 +02:00
Allan Odgaard
e49504a644 Remove unused member data 2013-05-06 14:51:32 +07:00
Allan Odgaard
45dc49eb29 Retire compiler workaround
Previously we could get a compiler error when using std::map<std::string, std::string>() as a default argument.
2013-03-16 17:49:49 +01:00
Allan Odgaard
0075b46c82 Introduce zeroIndentPattern
This will give the matched lines zero indent but without affecting the following lines.

Probably the only use-case for this is C preprocessor directives.
2013-03-13 12:38:12 +01:00
Allan Odgaard
d16f3bcc1e Use per-line indent patterns when estimating indent
This solves the problem where we need to estimate the current line’s indent, but the lines above it is a multi-line block comment. Previously we would fetch indent patterns based on the current scope, then find the first line above caret, for which the patterns can be used to estimate the indent. The problem is that the commented lines without comment markers would be treated as code, and used for the indent.

With the new approach, we can set different patterns for ‘comment.block’ (the C bundle already does this), which basically ignore all the lines, which will cause TextMate to use the code above the comment to estimate indent.

This commit closes textmate/c.tmbundle#3 and also closes textmate/php.tmbundle#24.
2013-03-13 12:38:12 +01:00
Allan Odgaard
b139ac5097 Change push_back → emplace_back (C++11)
This is mainly motivated by readability, so I only did a few select replacements.
2013-02-08 11:20:35 +01:00
Allan Odgaard
9eb4044fdb Switch to simpler regexp::search 2013-02-08 11:20:35 +01:00
Allan Odgaard
8b8e02d6da Add minimal tests for regex::search 2013-02-08 11:20:35 +01:00
Allan Odgaard
a8f1c9ef06 Add convenience API for searching std::string 2013-02-08 11:20:34 +01:00
Allan Odgaard
0ee796d7a7 Implement regexp::match_t::operator[]
This can be used to access the n’th capture (as a std::string). If there is no capture, NULL_STR is returned.
2013-02-08 11:20:34 +01:00