diff options
Diffstat (limited to 'src/nvim/regexp.c')
| -rw-r--r-- | src/nvim/regexp.c | 139 |
1 files changed, 71 insertions, 68 deletions
diff --git a/src/nvim/regexp.c b/src/nvim/regexp.c index 24dc86e034..bec3bc9648 100644 --- a/src/nvim/regexp.c +++ b/src/nvim/regexp.c @@ -461,15 +461,15 @@ static int toggle_Magic(int x) */ #define UCHARAT(p) ((int)*(char_u *)(p)) -/* Used for an error (down from) vim_regcomp(): give the error message, set - * rc_did_emsg and return NULL */ -#define EMSG_RET_NULL(m) return (EMSG(m), rc_did_emsg = true, (void *)NULL) -#define IEMSG_RET_NULL(m) return (IEMSG(m), rc_did_emsg = true, (void *)NULL) -#define EMSG_RET_FAIL(m) return (EMSG(m), rc_did_emsg = true, FAIL) +// Used for an error (down from) vim_regcomp(): give the error message, set +// rc_did_emsg and return NULL +#define EMSG_RET_NULL(m) return (emsg(m), rc_did_emsg = true, (void *)NULL) +#define IEMSG_RET_NULL(m) return (iemsg(m), rc_did_emsg = true, (void *)NULL) +#define EMSG_RET_FAIL(m) return (emsg(m), rc_did_emsg = true, FAIL) #define EMSG2_RET_NULL(m, c) \ - return (EMSG2((m), (c) ? "" : "\\"), rc_did_emsg = true, (void *)NULL) + return (semsg((m), (c) ? "" : "\\"), rc_did_emsg = true, (void *)NULL) #define EMSG2_RET_FAIL(m, c) \ - return (EMSG2((m), (c) ? "" : "\\"), rc_did_emsg = true, FAIL) + return (semsg((m), (c) ? "" : "\\"), rc_did_emsg = true, FAIL) #define EMSG_ONE_RET_NULL EMSG2_RET_NULL(_( \ "E369: invalid item in %s%%[]"), reg_magic == MAGIC_ALL) @@ -715,9 +715,9 @@ static int reg_magic; /* magicness of the pattern: */ #define MAGIC_ON 3 /* "\m" or 'magic' */ #define MAGIC_ALL 4 /* "\v" very magic */ -static int reg_string; /* matching with a string instead of a buffer - line */ -static int reg_strict; /* "[abc" is illegal */ +static int reg_string; // matching with a string instead of a buffer + // line +static int reg_strict; // "[abc" is illegal /* * META contains all characters that may be magic, except '^' and '$'. @@ -741,10 +741,10 @@ static char_u META_flags[] = { 1, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1 }; -static int curchr; /* currently parsed character */ -/* Previous character. Note: prevchr is sometimes -1 when we are not at the - * start, eg in /[ ^I]^ the pattern was never found even if it existed, - * because ^ was taken to be magic -- webb */ +static int curchr; // currently parsed character +// Previous character. Note: prevchr is sometimes -1 when we are not at the +// start, eg in /[ ^I]^ the pattern was never found even if it existed, +// because ^ was taken to be magic -- webb static int prevchr; static int prevprevchr; /* previous-previous character */ static int nextchr; /* used for ungetchr() */ @@ -783,7 +783,7 @@ static int get_equi_class(char_u **pp) char_u *p = *pp; if (p[1] == '=' && p[2] != NUL) { - l = (*mb_ptr2len)(p + 2); + l = utfc_ptr2len(p + 2); if (p[l + 2] == '=' && p[l + 3] == ']') { c = utf_ptr2char(p + 2); *pp += l + 4; @@ -1144,7 +1144,7 @@ static char_u *skip_anyof(char_u *p) if (*p == ']' || *p == '-') ++p; while (*p != NUL && *p != ']') { - if ((l = (*mb_ptr2len)(p)) > 1) { + if ((l = utfc_ptr2len(p)) > 1) { p += l; } else if (*p == '-') { p++; @@ -1237,7 +1237,7 @@ static int seen_endbrace(int refnum) } if (*p == NUL) { - EMSG(_("E65: Illegal back reference")); + emsg(_("E65: Illegal back reference")); rc_did_emsg = true; return false; } @@ -1767,7 +1767,7 @@ static char_u *regpiece(int *flagp) else sprintf((char *)IObuff, _("E62: Nested %s%c"), reg_magic == MAGIC_ALL ? "" : "\\", no_Magic(peekchr())); - EMSG_RET_NULL(IObuff); + EMSG_RET_NULL((char *)IObuff); } return ret; @@ -1929,7 +1929,7 @@ static char_u *regatom(int *flagp) sprintf((char *)IObuff, _("E64: %s%c follows nothing"), (c == '*' ? reg_magic >= MAGIC_ON : reg_magic == MAGIC_ALL) ? "" : "\\", c); - EMSG_RET_NULL(IObuff); + EMSG_RET_NULL((char *)IObuff); /* NOTREACHED */ case Magic('~'): /* previous substitute pattern */ @@ -2255,8 +2255,8 @@ collection: if (startc > endc) { EMSG_RET_NULL(_(e_reverse_range)); } - if ((*mb_char2len)(startc) > 1 - || (*mb_char2len)(endc) > 1) { + if (utf_char2len(startc) > 1 + || utf_char2len(endc) > 1) { // Limit to a range of 256 chars if (endc > startc + 256) { EMSG_RET_NULL(_(e_large_class)); @@ -2509,8 +2509,9 @@ do_multibyte: /* Need to get composing character too. */ for (;; ) { l = utf_ptr2len(regparse); - if (!UTF_COMPOSINGLIKE(regparse, regparse + l)) + if (!utf_composinglike(regparse, regparse + l)) { break; + } regmbc(utf_ptr2char(regparse)); skipchr(); } @@ -2835,21 +2836,22 @@ static int peekchr(void) curchr = Magic(curchr); break; case '*': - /* * is not magic as the very first character, eg "?*ptr", when - * after '^', eg "/^*ptr" and when after "\(", "\|", "\&". But - * "\(\*" is not magic, thus must be magic if "after_slash" */ + // * is not magic as the very first character, eg "?*ptr", when + // after '^', eg "/^*ptr" and when after "\(", "\|", "\&". But + // "\(\*" is not magic, thus must be magic if "after_slash" if (reg_magic >= MAGIC_ON && !at_start && !(prev_at_start && prevchr == Magic('^')) && (after_slash || (prevchr != Magic('(') && prevchr != Magic('&') - && prevchr != Magic('|')))) + && prevchr != Magic('|')))) { curchr = Magic('*'); + } break; case '^': - /* '^' is only magic as the very first character and if it's after - * "\(", "\|", "\&' or "\n" */ + // '^' is only magic as the very first character and if it's after + // "\(", "\|", "\&' or "\n" if (reg_magic >= MAGIC_OFF && (at_start || reg_magic == MAGIC_ALL @@ -2865,8 +2867,8 @@ static int peekchr(void) } break; case '$': - /* '$' is only magic as the very last char and if it's in front of - * either "\|", "\)", "\&", or "\n" */ + // '$' is only magic as the very last char and if it's in front of + // either "\|", "\)", "\&", or "\n" if (reg_magic >= MAGIC_OFF) { char_u *p = regparse + 1; bool is_magic_all = (reg_magic == MAGIC_ALL); @@ -3152,7 +3154,7 @@ static int read_limits(long *minval, long *maxval) if (*regparse != '}') { sprintf((char *)IObuff, _("E554: Syntax error in %s{...}"), reg_magic == MAGIC_ALL ? "" : "\\"); - EMSG_RET_FAIL(IObuff); + EMSG_RET_FAIL((char *)IObuff); } /* @@ -3482,7 +3484,7 @@ static long bt_regexec_both(char_u *line, /* Be paranoid... */ if (prog == NULL || line == NULL) { - IEMSG(_(e_null)); + iemsg(_(e_null)); goto theend; } @@ -3584,7 +3586,7 @@ static long bt_regexec_both(char_u *line, if (rex.line[col] == NUL) { break; } - col += (*mb_ptr2len)(rex.line + col); + col += utfc_ptr2len(rex.line + col); // Check for timeout once in a twenty times to avoid overhead. if (tm != NULL && ++tm_count == 20) { tm_count = 0; @@ -4126,7 +4128,7 @@ static bool regmatch( break; case PRINT: - if (!vim_isprintc(PTR2CHAR(rex.input))) { + if (!vim_isprintc(utf_ptr2char(rex.input))) { status = RA_NOMATCH; } else { ADVANCE_REGINPUT(); @@ -4134,7 +4136,7 @@ static bool regmatch( break; case SPRINT: - if (ascii_isdigit(*rex.input) || !vim_isprintc(PTR2CHAR(rex.input))) { + if (ascii_isdigit(*rex.input) || !vim_isprintc(utf_ptr2char(rex.input))) { status = RA_NOMATCH; } else { ADVANCE_REGINPUT(); @@ -4293,7 +4295,7 @@ static bool regmatch( // Check for following composing character, unless %C // follows (skips over all composing chars). if (status != RA_NOMATCH - && UTF_COMPOSINGLIKE(rex.input, rex.input + len) + && utf_composinglike(rex.input, rex.input + len) && !rex.reg_icombine && OP(next) != RE_COMPOSING) { // raaron: This code makes a composing character get @@ -4325,7 +4327,7 @@ static bool regmatch( const char_u *opnd = OPERAND(scan); // Safety check (just in case 'encoding' was changed since // compiling the program). - if ((len = (*mb_ptr2len)(opnd)) < 2) { + if ((len = utfc_ptr2len(opnd)) < 2) { status = RA_NOMATCH; break; } @@ -4729,7 +4731,7 @@ static bool regmatch( * follows. The code is below. Parameters are stored in * a regstar_T on the regstack. */ if ((long)((unsigned)regstack.ga_len >> 10) >= p_mmp) { - EMSG(_(e_maxmempat)); + emsg(_(e_maxmempat)); status = RA_FAIL; } else { ga_grow(®stack, sizeof(regstar_T)); @@ -4767,7 +4769,7 @@ static bool regmatch( case NOBEHIND: /* Need a bit of room to store extra positions. */ if ((long)((unsigned)regstack.ga_len >> 10) >= p_mmp) { - EMSG(_(e_maxmempat)); + emsg(_(e_maxmempat)); status = RA_FAIL; } else { ga_grow(®stack, sizeof(regbehind_T)); @@ -4815,7 +4817,7 @@ static bool regmatch( break; default: - IEMSG(_(e_re_corr)); + iemsg(_(e_re_corr)); #ifdef REGEXP_DEBUG printf("Illegal op code %d\n", op); #endif @@ -5173,7 +5175,7 @@ static bool regmatch( * We get here only if there's trouble -- normally "case END" is * the terminating point. */ - IEMSG(_(e_re_corr)); + iemsg(_(e_re_corr)); #ifdef REGEXP_DEBUG printf("Premature EOL\n"); #endif @@ -5195,7 +5197,7 @@ static regitem_T *regstack_push(regstate_T state, char_u *scan) regitem_T *rp; if ((long)((unsigned)regstack.ga_len >> 10) >= p_mmp) { - EMSG(_(e_maxmempat)); + emsg(_(e_maxmempat)); return NULL; } ga_grow(®stack, sizeof(regitem_T)); @@ -5268,7 +5270,7 @@ regrepeat ( case SIDENT: case SIDENT + ADD_NL: while (count < maxcount) { - if (vim_isIDc(PTR2CHAR(scan)) && (testval || !ascii_isdigit(*scan))) { + if (vim_isIDc(utf_ptr2char(scan)) && (testval || !ascii_isdigit(*scan))) { MB_PTR_ADV(scan); } else if (*scan == NUL) { if (!REG_MULTI || !WITH_NL(OP(p)) || rex.lnum > rex.reg_maxline @@ -5325,7 +5327,7 @@ regrepeat ( case SFNAME: case SFNAME + ADD_NL: while (count < maxcount) { - if (vim_isfilec(PTR2CHAR(scan)) && (testval || !ascii_isdigit(*scan))) { + if (vim_isfilec(utf_ptr2char(scan)) && (testval || !ascii_isdigit(*scan))) { MB_PTR_ADV(scan); } else if (*scan == NUL) { if (!REG_MULTI || !WITH_NL(OP(p)) || rex.lnum > rex.reg_maxline @@ -5363,7 +5365,7 @@ regrepeat ( if (got_int) { break; } - } else if (vim_isprintc(PTR2CHAR(scan)) == 1 + } else if (vim_isprintc(utf_ptr2char(scan)) == 1 && (testval || !ascii_isdigit(*scan))) { MB_PTR_ADV(scan); } else if (rex.reg_line_lbr && *scan == '\n' && WITH_NL(OP(p))) { @@ -5391,7 +5393,7 @@ do_class: if (got_int) { break; } - } else if ((l = (*mb_ptr2len)(scan)) > 1) { + } else if ((l = utfc_ptr2len(scan)) > 1) { if (testval != 0) { break; } @@ -5506,12 +5508,12 @@ do_class: /* Safety check (just in case 'encoding' was changed since * compiling the program). */ - if ((len = (*mb_ptr2len)(opnd)) > 1) { + if ((len = utfc_ptr2len(opnd)) > 1) { if (rex.reg_ic) { cf = utf_fold(utf_ptr2char(opnd)); } - while (count < maxcount && (*mb_ptr2len)(scan) >= len) { - for (i = 0; i < len; ++i) { + while (count < maxcount && utfc_ptr2len(scan) >= len) { + for (i = 0; i < len; i++) { if (opnd[i] != scan[i]) { break; } @@ -5580,7 +5582,7 @@ do_class: break; default: // Oh dear. Called inappropriately. - IEMSG(_(e_re_corr)); + iemsg(_(e_re_corr)); #ifdef REGEXP_DEBUG printf("Called regrepeat with op code %d\n", OP(p)); #endif @@ -5630,7 +5632,7 @@ static int prog_magic_wrong(void) } if (UCHARAT(((bt_regprog_T *)prog)->program) != REGMAGIC) { - EMSG(_(e_re_corr)); + emsg(_(e_re_corr)); return true; } return false; @@ -5811,8 +5813,8 @@ static int match_with_backref(linenr_T start_lnum, colnr_T start_col, linenr_T e if (bytelen != NULL) *bytelen = 0; for (;; ) { - /* Since getting one line may invalidate the other, need to make copy. - * Slow! */ + // Since getting one line may invalidate the other, need to make copy. + // Slow! if (rex.line != reg_tofree) { len = (int)STRLEN(rex.line); if (reg_tofree == NULL || len >= (int)reg_tofreelen) { @@ -6445,9 +6447,9 @@ static int cstrncmp(char_u *s1, char_u *s2, int *n) return result; } -/*************************************************************** -* regsub stuff * -***************************************************************/ +//////////////////////////////////////////////////////////////// +// regsub stuff // +//////////////////////////////////////////////////////////////// /* This stuff below really confuses cc on an SGI -- webb */ @@ -6512,9 +6514,10 @@ char_u *regtilde(char_u *source, int magic) memmove(tmpsub, newsub, (size_t)len); /* interpret tilde */ memmove(tmpsub + len, reg_prev_sub, (size_t)prevlen); - /* copy postfix */ - if (!magic) - ++p; /* back off \ */ + // copy postfix + if (!magic) { + p++; // back off backslash + } STRCPY(tmpsub + len + prevlen, p + 1); if (newsub != source) /* already allocated newsub */ @@ -6530,7 +6533,7 @@ char_u *regtilde(char_u *source, int magic) if (*p == '\\' && p[1]) { // skip escaped characters p++; } - p += (*mb_ptr2len)(p) - 1; + p += utfc_ptr2len(p) - 1; } } @@ -6687,7 +6690,7 @@ static int vim_regsub_both(char_u *source, typval_T *expr, char_u *dest, // Be paranoid... if ((source == NULL && expr == NULL) || dest == NULL) { - EMSG(_(e_null)); + emsg(_(e_null)); return 0; } if (prog_magic_wrong()) @@ -6941,7 +6944,7 @@ static int vim_regsub_both(char_u *source, typval_T *expr, char_u *dest, } } else if (*s == NUL) { // we hit NUL. if (copy) { - IEMSG(_(e_re_damg)); + iemsg(_(e_re_damg)); } goto exit; } else { @@ -7227,7 +7230,7 @@ regprog_T *vim_regcomp(char_u *expr_arg, int re_flags) regname[newengine]); #endif } else { - EMSG(_( + emsg(_( "E864: \\%#= can only be followed by 0, 1, or 2. The automatic engine will be used ")); regexp_engine = AUTOMATIC_ENGINE; } @@ -7261,7 +7264,7 @@ regprog_T *vim_regcomp(char_u *expr_arg, int re_flags) fprintf(f, "Syntax error in \"%s\"\n", expr); fclose(f); } else - EMSG2("(NFA) Could not open \"%s\" to write !!!", + semsg("(NFA) Could not open \"%s\" to write !!!", BT_REGEXP_DEBUG_LOG_NAME); } #endif @@ -7300,8 +7303,8 @@ static void report_re_switch(char_u *pat) { if (p_verbose > 0) { verbose_enter(); - MSG_PUTS(_("Switching to backtracking RE engine for pattern: ")); - MSG_PUTS(pat); + msg_puts(_("Switching to backtracking RE engine for pattern: ")); + msg_puts((char *)pat); verbose_leave(); } } @@ -7326,7 +7329,7 @@ static bool vim_regexec_string(regmatch_T *rmp, char_u *line, colnr_T col, // Cannot use the same prog recursively, it contains state. if (rmp->regprog->re_in_use) { - EMSG(_(e_recursive)); + emsg(_(e_recursive)); return false; } rmp->regprog->re_in_use = true; @@ -7423,7 +7426,7 @@ long vim_regexec_multi( // Cannot use the same prog recursively, it contains state. if (rmp->regprog->re_in_use) { - EMSG(_(e_recursive)); + emsg(_(e_recursive)); return false; } rmp->regprog->re_in_use = true; |