Mercurial > vim
changeset 5794:a63d0cd691dc v7.4.241
updated for version 7.4.241
Problem: The string returned by submatch() does not distinguish between a
NL from a line break and a NL that stands for a NUL character.
Solution: Add a second argument to return a list. (ZyX)
| author | Bram Moolenaar <bram@vim.org> |
|---|---|
| date | Wed, 02 Apr 2014 19:00:58 +0200 |
| parents | 3b08d14e08a3 |
| children | 64e4633f0751 |
| files | runtime/doc/eval.txt src/eval.c src/proto/regexp.pro src/regexp.c src/testdir/test79.in src/testdir/test79.ok src/testdir/test80.in src/testdir/test80.ok src/version.c |
| diffstat | 9 files changed, 128 insertions(+), 6 deletions(-) [+] |
line wrap: on
line diff
--- a/runtime/doc/eval.txt +++ b/runtime/doc/eval.txt @@ -1990,7 +1990,8 @@ strridx( {haystack}, {needle} [, {start} Number last index of {needle} in {haystack} strtrans( {expr}) String translate string to make it printable strwidth( {expr}) Number display cell length of the String {expr} -submatch( {nr}) String specific match in ":s" or substitute() +submatch( {nr}[, {list}]) String or List + specific match in ":s" or substitute() substitute( {expr}, {pat}, {sub}, {flags}) String all {pat} in {expr} replaced with {sub} synID( {lnum}, {col}, {trans}) Number syntax ID at {lnum} and {col} @@ -5797,12 +5798,23 @@ strwidth({expr}) *strwidth()* Ambiguous, this function's return value depends on 'ambiwidth'. Also see |strlen()|, |strdisplaywidth()| and |strchars()|. -submatch({nr}) *submatch()* +submatch({nr}[, {list}]) *submatch()* Only for an expression in a |:substitute| command or substitute() function. Returns the {nr}'th submatch of the matched text. When {nr} is 0 the whole matched text is returned. + Note that a NL in the string can stand for a line break of a + multi-line match or a NUL character in the text. Also see |sub-replace-expression|. + + If {list} is present and non-zero then submatch() returns + a list of strings, similar to |getline()| with two arguments. + NL characters in the text represent NUL characters in the + text. + Only returns more than one item for |:substitute|, inside + |substitute()| this list will always contain one or zero + items, since there are no real line breaks. + Example: > :s/\d\+/\=submatch(0) + 1/ < This finds the first number in the line and adds one to it.
--- a/src/eval.c +++ b/src/eval.c @@ -8129,7 +8129,7 @@ static struct fst {"strridx", 2, 3, f_strridx}, {"strtrans", 1, 1, f_strtrans}, {"strwidth", 1, 1, f_strwidth}, - {"submatch", 1, 1, f_submatch}, + {"submatch", 1, 2, f_submatch}, {"substitute", 4, 4, f_substitute}, {"synID", 3, 3, f_synID}, {"synIDattr", 2, 3, f_synIDattr}, @@ -17890,9 +17890,32 @@ f_submatch(argvars, rettv) typval_T *argvars; typval_T *rettv; { - rettv->v_type = VAR_STRING; - rettv->vval.v_string = - reg_submatch((int)get_tv_number_chk(&argvars[0], NULL)); + int error = FALSE; + char_u **match; + char_u **s; + listitem_T *li; + int no; + int retList = 0; + + no = (int)get_tv_number_chk(&argvars[0], &error); + if (error) + return; + error = FALSE; + if (argvars[1].v_type != VAR_UNKNOWN) + retList = get_tv_number_chk(&argvars[1], &error); + if (error) + return; + + if (retList == 0) + { + rettv->v_type = VAR_STRING; + rettv->vval.v_string = reg_submatch(no); + } + else + { + rettv->v_type = VAR_LIST; + rettv->vval.v_list = reg_submatch_list(no); + } } /*
--- a/src/proto/regexp.pro +++ b/src/proto/regexp.pro @@ -10,6 +10,7 @@ char_u *regtilde __ARGS((char_u *source, int vim_regsub __ARGS((regmatch_T *rmp, char_u *source, char_u *dest, int copy, int magic, int backslash)); int vim_regsub_multi __ARGS((regmmatch_T *rmp, linenr_T lnum, char_u *source, char_u *dest, int copy, int magic, int backslash)); char_u *reg_submatch __ARGS((int no)); +list_T *reg_submatch_list __ARGS((int no)); regprog_T *vim_regcomp __ARGS((char_u *expr_arg, int re_flags)); void vim_regfree __ARGS((regprog_T *prog)); int vim_regexec __ARGS((regmatch_T *rmp, char_u *line, colnr_T col));
--- a/src/regexp.c +++ b/src/regexp.c @@ -7897,6 +7897,85 @@ reg_submatch(no) return retval; } + +/* + * Used for the submatch() function with the optional non-zero argument: get + * the list of strings from the n'th submatch in allocated memory with NULs + * represented in NLs. + * Returns a list of allocated strings. Returns NULL when not in a ":s" + * command, for a non-existing submatch and for any error. + */ + list_T * +reg_submatch_list(no) + int no; +{ + char_u *s; + linenr_T slnum; + linenr_T elnum; + colnr_T scol; + colnr_T ecol; + int i; + list_T *list; + int error = FALSE; + + if (!can_f_submatch || no < 0) + return NULL; + + if (submatch_match == NULL) + { + slnum = submatch_mmatch->startpos[no].lnum; + elnum = submatch_mmatch->endpos[no].lnum; + if (slnum < 0 || elnum < 0) + return NULL; + + scol = submatch_mmatch->startpos[no].col; + ecol = submatch_mmatch->endpos[no].col; + + list = list_alloc(); + if (list == NULL) + return NULL; + + s = reg_getline_submatch(slnum) + scol; + if (slnum == elnum) + { + if (list_append_string(list, s, ecol - scol) == FAIL) + error = TRUE; + } + else + { + if (list_append_string(list, s, -1) == FAIL) + error = TRUE; + for (i = 1; i < elnum - slnum; i++) + { + s = reg_getline_submatch(slnum + i); + if (list_append_string(list, s, -1) == FAIL) + error = TRUE; + } + s = reg_getline_submatch(elnum); + if (list_append_string(list, s, ecol) == FAIL) + error = TRUE; + } + } + else + { + s = submatch_match->startp[no]; + if (s == NULL || submatch_match->endp[no] == NULL) + return NULL; + list = list_alloc(); + if (list == NULL) + return NULL; + if (list_append_string(list, s, + (int)(submatch_match->endp[no] - s)) == FAIL) + error = TRUE; + } + + if (error) + { + list_free(list, TRUE); + return NULL; + } + return list; +} #endif static regengine_T bt_regengine =
index f15ecc0f8dc07e351bf7ac6e1a8e1cba63db1fbb..56955c23182a8809b9aa233afbc8f6d1148d07c1 GIT binary patch literal 3123 zc%1E4OK;jh5av8mt~t9Xi6KqA<(;_VK)@uBG%X3Fhfax%G2mcq;zwGkYJYpztRJk? z3OMvuA+5e|cV<01-!eOW-Svct?Fh%NKRR{vrKxKAsk_uQLp6^D)f)3F`uC-yyRKnd z$AWDMGF@{$!}C2|f_}6Yom{A%Hp}HXn{|_8**$!{1Kjr;cm2L{-v`;X1$UvEru1Sg zK)G$LX5mhEebaLVdm?yruUbLoopeuFs8d5@P+QtZK!(8@fhXA|4RUnE&;`K|H8Ce) ze%pcIJ3PlTtYH-^SjG|-vA{NJkud`b`Ncq+V2CaW`qL@HR|E$$RA!8I2%4P^!(D<m zT~xY^T~oU`>T||!$o?Kf9|$@O-4cu@KN7SU`b2QVP>-O;&>cZ!=$@d)P@n8)7#h&e zqVErUG*KC_xd-}jG#_3b&`0HkBnkxMBTah)E}&T`mdcfC?d%*{pg<cg;R-s?g=;Wx zOjvwafOTtu{n3WyrzJf1p24|uz`b|D>w7SOhga%Z_OANj;=h8sD_Hzz3a<1!`m2Cg z7g3Bgq$Fa^fbH*zkrcsNOciDt^Y)56Vyq2y&4{r&fonuNp=-puD{hIgMD;8>*+`;* zS=n^9N;0So6-E~$(3%>ITVv1_9VYDwOfRP}yPAQ~G2p5D1oP|pDs?-jB`JitxPZRP zp9rEP(gA}(%qNN4V9wb$H`>$0uGzOX@6S6ZTLi>=v^RS=>-LwMkC0inzuMvmoI_`u zBXAzw%^ude{q-jM;2y5o-)y$GB76S2mjCs2-Lc@-*J2l>q(3<4*cr!a9IJAy!m%>P zN*pV4tPuTwc8=e33~QSElAPm2C{P)alce}riq}%Sn&OodFQ<4Z#fvFk2>91m7okF{ zd{W^r68IOnBatK7-`1o0q(+tJFj+mVN99F}>c(L@f@=8A(C1~#EB06mzbWw-Ck)_^ zvrUPK<T<M4=0df2vGP%ElhA5yZLQ#=gg;JyDqDGwVebHfXZy<#t*<>kQYkyJ?U5Qh zHF8#Hyx7K=cE5t|IJ6%gIl4M`g|QuUwKOx!pf~<(SY!K}E2UCeqbc2^NzKr-W@utF z!PA|A@TWkJ{?>CnCBv8z>4Xf!?Z_D9!h13ddunoW@a_4{s;!o6-ve2r-M28+Y4oa- I%f~_d21281DF6Tf
index bb30d140525facb55d0b400578271a9d90da5534..0f6ea4545106dbc331076866a7effa74ebcd0927 GIT binary patch literal 556 zc$}?}J9C0S6ooC5Gn4;76SlG8jZ_h2@~TYKh9wzAP|<+GV(D+s)rB-MUFMtN+`Yp& zXW=cao1LhYGE+oa3rF(+Ov3lcCr!g4R0NuCn3nChgMsVVmTBl3mc$8<D?F{Sc}D(% z*DZF$7g5|5D0?OLH+vjz52$#BI;kOg0@J3M;y9ZfyB6Jps>1)H<@8T2hKD(yW5M|X z39;l9o^p;98Lu*|h&2b*&~AgA7df_&Wo?AnMYIW*fU_U1{o+2G@ylz`ue|boQN;Y8 zHH-U7#9DEsI8tmWmK0NpAw`#>u{iG9Y`Q6t;J({9)KB<C#Db3p76&wpXy_|!#9}&x M@^I}L(hFk(-&MncWB>pF
--- a/src/testdir/test80.in +++ b/src/testdir/test80.in @@ -117,6 +117,7 @@ STARTTEST :set cpo& :$put =\"\n\nTEST_5:\" :$put =substitute('A123456789', 'A\(.\)\(.\)\(.\)\(.\)\(.\)\(.\)\(.\)\(.\)\(.\)', '\=submatch(0) . submatch(9) . submatch(8) . submatch(7) . submatch(6) . submatch(5) . submatch(4) . submatch(3) . submatch(2) . submatch(1)', '') +:$put =substitute('A123456789', 'A\(.\)\(.\)\(.\)\(.\)\(.\)\(.\)\(.\)\(.\)\(.\)', '\=string([submatch(0, 1), submatch(9, 1), submatch(8, 1), submatch(7, 1), submatch(6, 1), submatch(5, 1), submatch(4, 1), submatch(3, 1), submatch(2, 1), submatch(1, 1)])', '') /^TEST_6 ENDTEST @@ -142,6 +143,7 @@ STARTTEST :$put =\"\n\nTEST_7:\" :$put =substitute('A A', 'A.', '\=submatch(0)', '') :$put =substitute(\"B\nB\", 'B.', '\=submatch(0)', '') +:$put =substitute(\"B\nB\", 'B.', '\=string(submatch(0, 1))', '') :$put =substitute('-bb', '\zeb', 'a', 'g') :$put =substitute('-bb', '\ze', 'c', 'g') /^TEST_8