| 1 | #ifndef __GLOBISH_HH__␊ |
| 2 | #define __GLOBISH_HH__␊ |
| 3 | ␊ |
| 4 | // Copyright (C) 2005 Nathaniel Smith <njs@pobox.com>␊ |
| 5 | //␊ |
| 6 | // This program is made available under the GNU GPL version 2.0 or␊ |
| 7 | // greater. See the accompanying file COPYING for details.␊ |
| 8 | //␊ |
| 9 | // This program is distributed WITHOUT ANY WARRANTY; without even the␊ |
| 10 | // implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR␊ |
| 11 | // PURPOSE.␊ |
| 12 | ␊ |
| 13 | // a sort of glob-like pattern matcher, for use in specifying branch␊ |
| 14 | // collections for netsync. it is important that it not be too expensive to␊ |
| 15 | // match (as opposed to common regex engines, which can be exponential on␊ |
| 16 | // pathological patterns), because we must match branches against untrusted␊ |
| 17 | // patterns when doing netsync.␊ |
| 18 | ␊ |
| 19 | // all characters stand for themselves except␊ |
| 20 | //␊ |
| 21 | // \x matches x, even if x is one of the metacharacters␊ |
| 22 | // * matches zero or more characters of any kind (greedily)␊ |
| 23 | // ? matches any single character␊ |
| 24 | // [...] matches any single character that appears within the brackets␊ |
| 25 | // [^..] matches any single character that does _not_ appear␊ |
| 26 | // [!..] same as [^..]␊ |
| 27 | // {a,b,c} matches a or b or c (may be of arbitrary length, have arbitrary␊ |
| 28 | // number of alternations; nesting is allowed but only five deep)␊ |
| 29 | //␊ |
| 30 | // [\]] is how you put a ] in a character class␊ |
| 31 | // [\[] similarly (otherwise a syntax error)␊ |
| 32 | // [\\] similarly␊ |
| 33 | // [{}?*] within [] these stand for themselves␊ |
| 34 | //␊ |
| 35 | // \n matches n, not newline␊ |
| 36 | // \007 same as '007'␊ |
| 37 | //␊ |
| 38 | // to match, the _entire_ target must match the pattern; there is no scan␊ |
| 39 | // for a substring match, nor is a prefix match a match. the pattern is␊ |
| 40 | // expected to be utf8, and characters in the 0x00 - 0x1f range are not␊ |
| 41 | // permitted.␊ |
| 42 | //␊ |
| 43 | // as an extra special case, the empty string matches nothing, not even an␊ |
| 44 | // empty string. this hardly ever matters, but it's nice to have some way␊ |
| 45 | // to say "don't exclude anything", for instance.␊ |
| 46 | ␊ |
| 47 | #include "vector.hh"␊ |
| 48 | ␊ |
| 49 | struct arg_type;␊ |
| 50 | ␊ |
| 51 | struct globish␊ |
| 52 | {␊ |
| 53 | globish() : compiled_pattern() {}␊ |
| 54 | globish(char const * pat);␊ |
| 55 | globish(std::string const & pat);␊ |
| 56 | globish(std::vector<arg_type> const & pat);␊ |
| 57 | globish(std::vector<arg_type>::const_iterator const & beg,␊ |
| 58 | std::vector<arg_type>::const_iterator const & end);␊ |
| 59 | ␊ |
| 60 | std::string operator()(void) const;␊ |
| 61 | bool matches(std::string const & target) const;␊ |
| 62 | ␊ |
| 63 | private:␊ |
| 64 | std::string compiled_pattern;␊ |
| 65 | };␊ |
| 66 | ␊ |
| 67 | std::ostream & operator<<(std::ostream &, globish const &);␊ |
| 68 | template <> void dump(globish const &, std::string &);␊ |
| 69 | ␊ |
| 70 | // convenience functor for when you want to match all things␊ |
| 71 | // that _do_ match one glob but do _not_ match another␊ |
| 72 | struct globish_matcher␊ |
| 73 | {␊ |
| 74 | globish_matcher(globish const & incl, globish const & excl)␊ |
| 75 | : included(incl), excluded(excl) {}␊ |
| 76 | ␊ |
| 77 | bool operator()(std::string const & s)␊ |
| 78 | { return included.matches(s) && !excluded.matches(s); }␊ |
| 79 | ␊ |
| 80 | private:␊ |
| 81 | globish included;␊ |
| 82 | globish excluded;␊ |
| 83 | };␊ |
| 84 | ␊ |
| 85 | ␊ |
| 86 | // Local Variables:␊ |
| 87 | // mode: C++␊ |
| 88 | // fill-column: 76␊ |
| 89 | // c-file-style: "gnu"␊ |
| 90 | // indent-tabs-mode: nil␊ |
| 91 | // End:␊ |
| 92 | // vim: et:sw=2:sts=2:ts=2:cino=>2s,{s,\:s,+s,t0,g0,^-2,e-2,n-2,p2s,(0,=s:␊ |
| 93 | ␊ |
| 94 | #endif␊ |
