Date: Sun, 10 Aug 2008 17:41:25 GMT From: Gabor Kovesdan <gabor@FreeBSD.org> To: Perforce Change Reviews <perforce@freebsd.org> Subject: PERFORCE change 147093 for review Message-ID: <200808101741.m7AHfPgs040374@repoman.freebsd.org>
next in thread | raw e-mail | index | archive | help
http://perforce.freebsd.org/chv.cgi?CH=147093 Change 147093 by gabor@gabor_server on 2008/08/10 17:40:42 - More C-ish variable names - style(9) Affected files ... .. //depot/projects/soc2008/gabor_textproc/grep/fastgrep.c#6 edit .. //depot/projects/soc2008/gabor_textproc/grep/grep.h#41 edit Differences ... ==== //depot/projects/soc2008/gabor_textproc/grep/fastgrep.c#6 (text+ko) ==== @@ -49,35 +49,35 @@ int i; /* Initialize. */ - fg->patternLen = strlen(pattern); + fg->len = strlen(pattern); fg->bol = 0; fg->eol = 0; - fg->reversedSearch = 0; + fg->reversed = 0; /* * Make a copy and upper case it for later if in -i mode, * else just copy the pointer. */ if (iflag) { - fg->pattern = grep_malloc(fg->patternLen + 1); - for (i = 0; i < fg->patternLen; i++) + fg->pattern = grep_malloc(fg->len + 1); + for (i = 0; i < fg->len; i++) fg->pattern[i] = toupper(pattern[i]); - fg->pattern[fg->patternLen] = '\0'; + fg->pattern[fg->len] = '\0'; } else fg->pattern = (unsigned char *)pattern; /* really const */ /* Preprocess pattern. */ for (i = 0; i <= UCHAR_MAX; i++) - fg->qsBc[i] = fg->patternLen; - for (i = 1; i < fg->patternLen; i++) { - fg->qsBc[fg->pattern[i]] = fg->patternLen - i; + fg->qsBc[i] = fg->len; + for (i = 1; i < fg->len; i++) { + fg->qsBc[fg->pattern[i]] = fg->len - i; /* * If case is ignored, make the jump apply to both upper and * lower cased characters. As the pattern is stored in upper * case, apply the same to the lower case equivalents. */ if (iflag) - fg->qsBc[tolower(fg->pattern[i])] = fg->patternLen - i; + fg->qsBc[tolower(fg->pattern[i])] = fg->len - i; } } @@ -97,29 +97,30 @@ int lastHalfDot = 0; /* Initialize. */ - fg->patternLen = strlen(pattern); + fg->len = strlen(pattern); fg->bol = 0; fg->eol = 0; - fg->reversedSearch = 0; + fg->reversed = 0; /* Remove end-of-line character ('$'). */ - if (pattern[fg->patternLen - 1] == '$') { + if (pattern[fg->len - 1] == '$') { eol++; fg->eol = 1; - fg->patternLen--; + fg->len--; } /* Remove beginning-of-line character ('^'). */ if (pattern[0] == '^') { bol++; fg->bol = 1; - fg->patternLen--; + fg->len--; } - if (fg->patternLen >= 14 && + if (fg->len >= 14 && strncmp(pattern + fg->bol, "[[:<:]]", 7) == 0 && - strncmp(pattern + fg->bol + fg->patternLen - 7, "[[:>:]]", 7) == 0) { - fg->patternLen -= 14; + strncmp(pattern + fg->bol + fg->len - 7, "[[:>:]]", 7) == 0) { + fg->len -= 14; + /* Word boundary is handled separately in util.c */ wflag = 1; } @@ -128,12 +129,12 @@ * match character classes at the beginning and ending of the * string respectively. */ - fg->pattern = grep_malloc(fg->patternLen + 1); - memcpy(fg->pattern, pattern + bol + wflag, fg->patternLen); - fg->pattern[fg->patternLen] = '\0'; + fg->pattern = grep_malloc(fg->len + 1); + memcpy(fg->pattern, pattern + bol + wflag, fg->len); + fg->pattern[fg->len] = '\0'; /* Look for ways to cheat...er...avoid the full regex engine. */ - for (i = 0; i < fg->patternLen; i++) + for (i = 0; i < fg->len; i++) { /* Can still cheat? */ if ((isalnum(fg->pattern[i])) || isspace(fg->pattern[i]) || @@ -145,7 +146,7 @@ fg->pattern[i] = toupper(fg->pattern[i]); } else if (fg->pattern[i] == '.') { hasDot = i; - if (i < fg->patternLen / 2) { + if (i < fg->len / 2) { if (firstHalfDot < 0) /* Closest dot to the beginning */ firstHalfDot = i; @@ -169,11 +170,11 @@ */ if ((!(lflag || cflag)) && ((!(bol || eol)) && ((lastHalfDot) && ((firstHalfDot < 0) || - ((fg->patternLen - (lastHalfDot + 1)) < firstHalfDot)))) && !oflag && !color) { - fg->reversedSearch = 1; - hasDot = fg->patternLen - (firstHalfDot < 0 ? + ((fg->len - (lastHalfDot + 1)) < firstHalfDot)))) && !oflag && !color) { + fg->reversed = 1; + hasDot = fg->len - (firstHalfDot < 0 ? firstLastHalfDot : firstHalfDot) - 1; - grep_revstr(fg->pattern, fg->patternLen); + grep_revstr(fg->pattern, fg->len); } /* @@ -196,105 +197,105 @@ */ /* Adjust the shift based on location of the last dot ('.'). */ - shiftPatternLen = fg->patternLen - hasDot; + shiftPatternLen = fg->len - hasDot; /* Preprocess pattern. */ for (i = 0; i <= UCHAR_MAX; i++) fg->qsBc[i] = shiftPatternLen; - for (i = hasDot + 1; i < fg->patternLen; i++) { - fg->qsBc[fg->pattern[i]] = fg->patternLen - i; + for (i = hasDot + 1; i < fg->len; i++) { + fg->qsBc[fg->pattern[i]] = fg->len - i; /* * If case is ignored, make the jump apply to both upper and * lower cased characters. As the pattern is stored in upper * case, apply the same to the lower case equivalents. */ if (iflag) - fg->qsBc[tolower(fg->pattern[i])] = fg->patternLen - i; + fg->qsBc[tolower(fg->pattern[i])] = fg->len - i; } /* * Put pattern back to normal after pre-processing to allow for easy * comparisons later. */ - if (fg->reversedSearch) - grep_revstr(fg->pattern, fg->patternLen); + if (fg->reversed) + grep_revstr(fg->pattern, fg->len); return (0); } int -grep_search(fastgrep_t *fg, unsigned char *data, size_t dataLen, regmatch_t *pmatch) +grep_search(fastgrep_t *fg, unsigned char *data, size_t len, regmatch_t *pmatch) { - int j; - int rtrnVal = REG_NOMATCH; + int j; + int ret = REG_NOMATCH; - if (pmatch->rm_so == dataLen) - return (rtrnVal); + if (pmatch->rm_so == len) + return (ret); if (fg->bol && pmatch->rm_so != 0) { - pmatch->rm_so = dataLen; - pmatch->rm_eo = dataLen; - return (rtrnVal); + pmatch->rm_so = len; + pmatch->rm_eo = len; + return (ret); } /* No point in going farther if we do not have enough data. */ - if (dataLen < fg->patternLen) - return (rtrnVal); + if (len < fg->len) + return (ret); /* Only try once at the beginning or ending of the line. */ if (fg->bol || fg->eol) { /* Simple text comparison. */ /* Verify data is >= pattern length before searching on it. */ - if (dataLen >= fg->patternLen) { + if (len >= fg->len) { /* Determine where in data to start search at. */ if (fg->eol) - j = dataLen - fg->patternLen; + j = len - fg->len; else j = 0; - if (!((fg->bol && fg->eol) && (dataLen != fg->patternLen))) + if (!((fg->bol && fg->eol) && (len != fg->len))) if (grep_cmp(fg->pattern, data + j, - fg->patternLen) == -1) { + fg->len) == -1) { pmatch->rm_so = j; - pmatch->rm_eo = j + fg->patternLen; - rtrnVal = 0; + pmatch->rm_eo = j + fg->len; + ret = 0; } } - } else if (fg->reversedSearch) { + } else if (fg->reversed) { /* Quick Search algorithm. */ - j = dataLen; + j = len; do { - if (grep_cmp(fg->pattern, data + j - fg->patternLen, - fg->patternLen) == -1) { - pmatch->rm_so = j - fg->patternLen; + if (grep_cmp(fg->pattern, data + j - fg->len, + fg->len) == -1) { + pmatch->rm_so = j - fg->len; pmatch->rm_eo = j; - rtrnVal = 0; + ret = 0; break; } /* Shift if within bounds, otherwise, we are done. */ - if (j == fg->patternLen) + if (j == fg->len) break; - j -= fg->qsBc[data[j - fg->patternLen - 1]]; - } while (j >= fg->patternLen); + j -= fg->qsBc[data[j - fg->len - 1]]; + } while (j >= fg->len); } else { /* Quick Search algorithm. */ j = pmatch->rm_so; do { - if (grep_cmp(fg->pattern, data + j, fg->patternLen) == -1) { + if (grep_cmp(fg->pattern, data + j, fg->len) == -1) { pmatch->rm_so = j; - pmatch->rm_eo = j + fg->patternLen; - rtrnVal = 0; + pmatch->rm_eo = j + fg->len; + ret = 0; break; } /* Shift if within bounds, otherwise, we are done. */ - if (j + fg->patternLen == dataLen) + if (j + fg->len == len) break; else - j += fg->qsBc[data[j + fg->patternLen]]; - } while (j <= (dataLen - fg->patternLen)); + j += fg->qsBc[data[j + fg->len]]; + } while (j <= (len - fg->len)); } - return (rtrnVal); + return (ret); } /* @@ -304,7 +305,7 @@ static int grep_cmp(const unsigned char *pattern, const unsigned char *data, size_t len) { - int i; + int i; for (i = 0; i < len; i++) { if (((pattern[i] == data[i]) || ((grepbehave != GREP_FIXED) && pattern[i] == '.')) @@ -319,8 +320,8 @@ static void grep_revstr(unsigned char *str, int len) { - int i; - char c; + int i; + char c; for (i = 0; i < len / 2; i++) { c = str[i]; ==== //depot/projects/soc2008/gabor_textproc/grep/grep.h#41 (text+ko) ==== @@ -93,13 +93,12 @@ typedef struct { unsigned char *pattern; - int patternLen; + int len; int qsBc[UCHAR_MAX + 1]; /* flags */ int bol; int eol; - int wmatch; - int reversedSearch; + int reversed; } fastgrep_t; /* Flags passed to regcomp() and regexec() */
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200808101741.m7AHfPgs040374>