Mercurial > octave-nkf
changeset 20246:f2bc7d23295d
Add special hex/octal escape sequence processing for regexp.
* NEWS: Announce change done for Matlab compatibility.
* regexp.cc (do_regexp_ptn_string_escapes): Add case for parsing '\oNNN' or
'\o{NNN}' octal sequences. PCRE already handles the hex case.
* regexp.cc (do_regexp_rep_string_escapes): Add case for parsing '\oNNN' or
'\o{NNN}' octal sequences. Add case for parsing '\xNN' or '\x{NN}'.
| author | Rik <rik@octave.org> |
|---|---|
| date | Thu, 21 May 2015 10:22:27 -0700 |
| parents | 29eb47fe8e8c |
| children | 561af1ab6099 |
| files | NEWS libinterp/corefcn/regexp.cc |
| diffstat | 2 files changed, 106 insertions(+), 14 deletions(-) [+] |
line wrap: on
line diff
--- a/NEWS Wed May 20 17:18:41 2015 -0700 +++ b/NEWS Thu May 21 10:22:27 2015 -0700 @@ -6,6 +6,12 @@ The *printf family of functions now supports octal and hex escape sequences in single-quoted strings for Matlab compatibility. + ** Special octal and hex escape sequences for the pattern and replacement + strings in regular expressions are now interpreted for Matlab compatibility. + + octal: '\oNNN' or '\o{NNN}' + hex : '\xNN' or '\x{NN}' + ** mkfifo now interprets the MODE argument as an octal, not decimal, integer. This is consistent with the equivalent shell command.
--- a/libinterp/corefcn/regexp.cc Wed May 20 17:18:41 2015 -0700 +++ b/libinterp/corefcn/regexp.cc Thu May 21 10:22:27 2015 -0700 @@ -77,13 +77,38 @@ retval[++i] = 'b'; break; -#if 0 -// FIXME: To be complete, we need to handle \oN, \o{N}. -// The PCRE library already handles \N where N -// is an octal number. New code needs to merely -// replace \oN or \o{N} with \N. - case 'o': // octal number -#endif + case 'o': // octal input + { + bool bad_esc_seq = (j+1 >= len); + + bool brace = false; + if (! bad_esc_seq && s[++j] == '{') + { + brace = true; + j++; + } + + int tmpi = 0; + size_t k; + for (k = j; k < std::min (j+3+brace, len); k++) + { + int digit = s[k] - '0'; + if (digit < 0 || digit > 7) + break; + tmpi <<= 3; + tmpi += digit; + } + if (bad_esc_seq || (brace && s[k++] != '}')) + { + bad_esc_seq = true; + tmpi = 0; + warning ("malformed octal escape sequence '\\o' --\ + converting to '\\0'"); + } + retval[i] = tmpi; + j = k - 1; + break; + } default: // pass escape sequence through retval[i] = '\\'; @@ -150,14 +175,75 @@ retval[i] = '\v'; break; -#if 0 -// FIXME: to be complete, we need to handle \oN, \o{N}, \xN, and -// \x{N}. Hex digits may be upper or lower case. Brackets are -// optional, so \x5Bz is the same as \x{5B}z. + case 'o': // octal input + { + bool bad_esc_seq = (j+1 >= len); + + bool brace = false; + if (! bad_esc_seq && s[++j] == '{') + { + brace = true; + j++; + } + + int tmpi = 0; + size_t k; + for (k = j; k < std::min (j+3+brace, len); k++) + { + int digit = s[k] - '0'; + if (digit < 0 || digit > 7) + break; + tmpi <<= 3; + tmpi += digit; + } + if (bad_esc_seq || (brace && s[k++] != '}')) + { + warning ("malformed octal escape sequence '\\o' --\ + converting to '\\0'"); + tmpi = 0; + } + retval[i] = tmpi; + j = k - 1; + break; + } - case 'o': // octal number - case 'x': // hex number -#endif + case 'x': // hex input + { + bool bad_esc_seq = (j+1 >= len); + + bool brace = false; + if (! bad_esc_seq && s[++j] == '{') + { + brace = true; + j++; + } + + int tmpi = 0; + size_t k; + for (k = j; k < std::min (j+2+brace, len); k++) + { + if (! isxdigit (s[k])) + break; + + tmpi <<= 4; + int digit = s[k]; + if (digit >= 'a') + tmpi += digit - 'a' + 10; + else if (digit >= 'A') + tmpi += digit - 'A' + 10; + else + tmpi += digit - '0'; + } + if (bad_esc_seq || (brace && s[k++] != '}')) + { + warning ("malformed hex escape sequence '\\x' --\ + converting to '\\0'"); + tmpi = 0; + } + retval[i] = tmpi; + j = k - 1; + break; + } default: // pass escape sequence through retval[i] = '\\';