aboutsummaryrefslogtreecommitdiff
path: root/src/nvim/search.c
diff options
context:
space:
mode:
Diffstat (limited to 'src/nvim/search.c')
-rw-r--r--src/nvim/search.c898
1 files changed, 724 insertions, 174 deletions
diff --git a/src/nvim/search.c b/src/nvim/search.c
index 906c9a6f47..f3061b4dc4 100644
--- a/src/nvim/search.c
+++ b/src/nvim/search.c
@@ -13,8 +13,8 @@
#include "nvim/ascii.h"
#include "nvim/buffer.h"
-#include "nvim/charset.h"
#include "nvim/change.h"
+#include "nvim/charset.h"
#include "nvim/cursor.h"
#include "nvim/edit.h"
#include "nvim/eval.h"
@@ -26,6 +26,7 @@
#include "nvim/func_attr.h"
#include "nvim/getchar.h"
#include "nvim/indent.h"
+#include "nvim/indent_c.h"
#include "nvim/main.h"
#include "nvim/mark.h"
#include "nvim/mbyte.h"
@@ -175,7 +176,7 @@ int search_regcomp(char_u *pat, int pat_save, int pat_use, int options, regmmatc
* Save the currently used pattern in the appropriate place,
* unless the pattern should not be remembered.
*/
- if (!(options & SEARCH_KEEP) && !cmdmod.keeppatterns) {
+ if (!(options & SEARCH_KEEP) && (cmdmod.cmod_flags & CMOD_KEEPPATTERNS) == 0) {
// search or global command
if (pat_save == RE_SEARCH || pat_save == RE_BOTH) {
save_re_pat(RE_SEARCH, pat, magic);
@@ -188,7 +189,7 @@ int search_regcomp(char_u *pat, int pat_save, int pat_use, int options, regmmatc
regmatch->rmm_ic = ignorecase(pat);
regmatch->rmm_maxcol = 0;
- regmatch->regprog = vim_regcomp(pat, magic ? RE_MAGIC : 0);
+ regmatch->regprog = vim_regcomp((char *)pat, magic ? RE_MAGIC : 0);
if (regmatch->regprog == NULL) {
return FAIL;
}
@@ -203,31 +204,6 @@ char_u *get_search_pat(void)
return mr_pattern;
}
-/*
- * Reverse text into allocated memory.
- * Returns the allocated string.
- *
- * TODO(philix): move reverse_text() to strings.c
- */
-char_u *reverse_text(char_u *s) FUNC_ATTR_NONNULL_RET
-{
- /*
- * Reverse the pattern.
- */
- size_t len = STRLEN(s);
- char_u *rev = xmalloc(len + 1);
- size_t rev_i = len;
- for (size_t s_i = 0; s_i < len; s_i++) {
- const int mb_len = utfc_ptr2len(s + s_i);
- rev_i -= mb_len;
- memmove(rev + rev_i, s + s_i, mb_len);
- s_i += mb_len - 1;
- }
- rev[len] = NUL;
-
- return rev;
-}
-
void save_re_pat(int idx, char_u *pat, int magic)
{
if (spats[idx].pat != pat) {
@@ -310,6 +286,8 @@ static struct spat saved_last_search_spat;
static int did_save_last_search_spat = 0;
static int saved_last_idx = 0;
static bool saved_no_hlsearch = false;
+static colnr_T saved_search_match_endcol;
+static linenr_T saved_search_match_lines;
/// Save and restore the search pattern for incremental highlight search
/// feature.
@@ -352,6 +330,21 @@ void restore_last_search_pattern(void)
set_no_hlsearch(saved_no_hlsearch);
}
+/// Save and restore the incsearch highlighting variables.
+/// This is required so that calling searchcount() at does not invalidate the
+/// incsearch highlighting.
+static void save_incsearch_state(void)
+{
+ saved_search_match_endcol = search_match_endcol;
+ saved_search_match_lines = search_match_lines;
+}
+
+static void restore_incsearch_state(void)
+{
+ search_match_endcol = saved_search_match_endcol;
+ search_match_lines = saved_search_match_lines;
+}
+
char_u *last_search_pattern(void)
{
return spats[RE_SEARCH].pat;
@@ -387,10 +380,10 @@ bool pat_has_uppercase(char_u *pat)
char_u *p = pat;
while (*p != NUL) {
- const int l = utfc_ptr2len(p);
+ const int l = utfc_ptr2len((char *)p);
if (l > 1) {
- if (mb_isupper(utf_ptr2char(p))) {
+ if (mb_isupper(utf_ptr2char((char *)p))) {
return true;
}
p += l;
@@ -521,7 +514,7 @@ void last_pat_prog(regmmatch_T *regmatch)
--emsg_off;
}
-/// lowest level search function.
+/// Lowest level search function.
/// Search for 'count'th occurrence of pattern "pat" in direction "dir".
/// Start at position "pos" and return the found position in "pos".
///
@@ -563,7 +556,7 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
long nmatched;
int submatch = 0;
bool first_match = true;
- int save_called_emsg = called_emsg;
+ const int called_emsg_before = called_emsg;
bool break_loop = false;
linenr_T stop_lnum = 0; // stop after this line number when != 0
proftime_T *tm = NULL; // timeout limit or NULL
@@ -586,7 +579,6 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
/*
* find the string
*/
- called_emsg = FALSE;
do { // loop for count
// When not accepting a match at the start position set "extra_col" to a
// non-zero value. Don't do that when starting at MAXCOL, since MAXCOL + 1
@@ -601,7 +593,7 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
if ((int)STRLEN(ptr) <= pos->col) {
start_char_len = 1;
} else {
- start_char_len = utfc_ptr2len(ptr + pos->col);
+ start_char_len = utfc_ptr2len((char *)ptr + pos->col);
}
} else {
start_char_len = 1;
@@ -658,7 +650,7 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
break;
}
// Abort searching on an error (e.g., out of stack).
- if (called_emsg || (timed_out != NULL && *timed_out)) {
+ if (called_emsg > called_emsg_before || (timed_out != NULL && *timed_out)) {
break;
}
if (nmatched > 0) {
@@ -713,7 +705,7 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
}
if (matchcol == matchpos.col && ptr[matchcol] != NUL) {
- matchcol += utfc_ptr2len(ptr + matchcol);
+ matchcol += utfc_ptr2len((char *)ptr + matchcol);
}
if (matchcol == 0 && (options & SEARCH_START)) {
@@ -798,7 +790,7 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
// for empty match: advance one char
if (matchcol == matchpos.col
&& ptr[matchcol] != NUL) {
- matchcol += utfc_ptr2len(ptr + matchcol);
+ matchcol += utfc_ptr2len((char *)ptr + matchcol);
}
} else {
// Stop when the match is in a next line.
@@ -807,7 +799,7 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
}
matchcol = matchpos.col;
if (ptr[matchcol] != NUL) {
- matchcol += utfc_ptr2len(ptr + matchcol);
+ matchcol += utfc_ptr2len((char *)ptr + matchcol);
}
}
if (ptr[matchcol] == NUL
@@ -915,7 +907,8 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
// Stop the search if wrapscan isn't set, "stop_lnum" is
// specified, after an interrupt, after a match and after looping
// twice.
- if (!p_ws || stop_lnum != 0 || got_int || called_emsg
+ if (!p_ws || stop_lnum != 0 || got_int
+ || called_emsg > called_emsg_before
|| (timed_out != NULL && *timed_out)
|| break_loop
|| found || loop) {
@@ -934,14 +927,13 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
lnum = 1;
}
if (!shortmess(SHM_SEARCH) && (options & SEARCH_MSG)) {
- give_warning((char_u *)_(dir == BACKWARD
- ? top_bot_msg : bot_top_msg), true);
+ give_warning(_(dir == BACKWARD ? top_bot_msg : bot_top_msg), true);
}
if (extra_arg != NULL) {
extra_arg->sa_wrapped = true;
}
}
- if (got_int || called_emsg
+ if (got_int || called_emsg > called_emsg_before
|| (timed_out != NULL && *timed_out)
|| break_loop) {
break;
@@ -950,8 +942,6 @@ int searchit(win_T *win, buf_T *buf, pos_T *pos, pos_T *end_pos, Direction dir,
vim_regfree(regmatch.regprog);
- called_emsg |= save_called_emsg;
-
if (!found) { // did not find it
if (got_int) {
emsg(_(e_interr));
@@ -1054,7 +1044,7 @@ int do_search(oparg_T *oap, int dirc, int search_delim, char_u *pat, long count,
* A line offset is not remembered, this is vi compatible.
*/
if (spats[0].off.line && vim_strchr(p_cpo, CPO_LINEOFF) != NULL) {
- spats[0].off.line = FALSE;
+ spats[0].off.line = false;
spats[0].off.off = 0;
}
@@ -1070,7 +1060,7 @@ int do_search(oparg_T *oap, int dirc, int search_delim, char_u *pat, long count,
* Find out the direction of the search.
*/
if (dirc == 0) {
- dirc = spats[0].off.dir;
+ dirc = (char_u)spats[0].off.dir;
} else {
spats[0].off.dir = dirc;
set_vv_searchforward();
@@ -1247,7 +1237,7 @@ int do_search(oparg_T *oap, int dirc, int search_delim, char_u *pat, long count,
// empty for the search_stat feature.
if (!cmd_silent) {
msgbuf[0] = dirc;
- if (utf_iscomposing(utf_ptr2char(p))) {
+ if (utf_iscomposing(utf_ptr2char((char *)p))) {
// Use a space to draw the composing char on.
msgbuf[1] = ' ';
memmove(msgbuf + 2, p, STRLEN(p));
@@ -1285,7 +1275,7 @@ int do_search(oparg_T *oap, int dirc, int search_delim, char_u *pat, long count,
memset(msgbuf + pat_len, ' ', r - msgbuf);
}
}
- msg_outtrans(msgbuf);
+ msg_outtrans((char *)msgbuf);
msg_clr_eos();
msg_check();
@@ -1444,7 +1434,7 @@ int do_search(oparg_T *oap, int dirc, int search_delim, char_u *pat, long count,
curwin->w_set_curswant = TRUE;
end_do_search:
- if ((options & SEARCH_KEEP) || cmdmod.keeppatterns) {
+ if ((options & SEARCH_KEEP) || (cmdmod.cmod_flags & CMOD_KEEPPATTERNS)) {
spats[0].off = old_off;
}
xfree(msgbuf);
@@ -1476,7 +1466,7 @@ int search_for_exact_line(buf_T *buf, pos_T *pos, Direction dir, char_u *pat)
if (p_ws) {
pos->lnum = buf->b_ml.ml_line_count;
if (!shortmess(SHM_SEARCH)) {
- give_warning((char_u *)_(top_bot_msg), true);
+ give_warning(_(top_bot_msg), true);
}
} else {
pos->lnum = 1;
@@ -1486,7 +1476,7 @@ int search_for_exact_line(buf_T *buf, pos_T *pos, Direction dir, char_u *pat)
if (p_ws) {
pos->lnum = 1;
if (!shortmess(SHM_SEARCH)) {
- give_warning((char_u *)_(bot_top_msg), true);
+ give_warning(_(bot_top_msg), true);
}
} else {
pos->lnum = 1;
@@ -1500,7 +1490,7 @@ int search_for_exact_line(buf_T *buf, pos_T *pos, Direction dir, char_u *pat)
start = pos->lnum;
}
ptr = ml_get_buf(buf, pos->lnum, false);
- p = skipwhite(ptr);
+ p = (char_u *)skipwhite((char *)ptr);
pos->col = (colnr_T)(p - ptr);
// when adding lines the matching line may be empty but it is not
@@ -1548,13 +1538,13 @@ int searchc(cmdarg_T *cap, int t_cmd)
*lastc = c;
set_csearch_direction(dir);
set_csearch_until(t_cmd);
- lastc_bytelen = utf_char2bytes(c, lastc_bytes);
+ lastc_bytelen = utf_char2bytes(c, (char *)lastc_bytes);
if (cap->ncharC1 != 0) {
lastc_bytelen += utf_char2bytes(cap->ncharC1,
- lastc_bytes + lastc_bytelen);
+ (char *)lastc_bytes + lastc_bytelen);
if (cap->ncharC2 != 0) {
lastc_bytelen += utf_char2bytes(cap->ncharC2,
- lastc_bytes + lastc_bytelen);
+ (char *)lastc_bytes + lastc_bytelen);
}
}
}
@@ -1592,7 +1582,7 @@ int searchc(cmdarg_T *cap, int t_cmd)
while (count--) {
for (;;) {
if (dir > 0) {
- col += utfc_ptr2len(p + col);
+ col += utfc_ptr2len((char *)p + col);
if (col >= len) {
return FAIL;
}
@@ -1708,31 +1698,31 @@ static void find_mps_values(int *initc, int *findc, bool *backwards, bool switch
char_u *ptr = curbuf->b_p_mps;
while (*ptr != NUL) {
- if (utf_ptr2char(ptr) == *initc) {
+ if (utf_ptr2char((char *)ptr) == *initc) {
if (switchit) {
*findc = *initc;
- *initc = utf_ptr2char(ptr + utfc_ptr2len(ptr) + 1);
+ *initc = utf_ptr2char((char *)ptr + utfc_ptr2len((char *)ptr) + 1);
*backwards = true;
} else {
- *findc = utf_ptr2char(ptr + utfc_ptr2len(ptr) + 1);
+ *findc = utf_ptr2char((char *)ptr + utfc_ptr2len((char *)ptr) + 1);
*backwards = false;
}
return;
}
char_u *prev = ptr;
- ptr += utfc_ptr2len(ptr) + 1;
- if (utf_ptr2char(ptr) == *initc) {
+ ptr += utfc_ptr2len((char *)ptr) + 1;
+ if (utf_ptr2char((char *)ptr) == *initc) {
if (switchit) {
*findc = *initc;
- *initc = utf_ptr2char(prev);
+ *initc = utf_ptr2char((char *)prev);
*backwards = false;
} else {
- *findc = utf_ptr2char(prev);
+ *findc = utf_ptr2char((char *)prev);
*backwards = true;
}
return;
}
- ptr += utfc_ptr2len(ptr);
+ ptr += utfc_ptr2len((char *)ptr);
if (*ptr == ',') {
ptr++;
}
@@ -1834,9 +1824,9 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
*/
if (!cpo_match) {
// Are we before or at #if, #else etc.?
- ptr = skipwhite(linep);
+ ptr = (char_u *)skipwhite((char *)linep);
if (*ptr == '#' && pos.col <= (colnr_T)(ptr - linep)) {
- ptr = skipwhite(ptr + 1);
+ ptr = (char_u *)skipwhite((char *)ptr + 1);
if (STRNCMP(ptr, "if", 2) == 0
|| STRNCMP(ptr, "endif", 5) == 0
|| STRNCMP(ptr, "el", 2) == 0) {
@@ -1879,7 +1869,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
--pos.col;
}
for (;;) {
- initc = utf_ptr2char(linep + pos.col);
+ initc = utf_ptr2char((char *)linep + pos.col);
if (initc == NUL) {
break;
}
@@ -1888,11 +1878,11 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
if (findc) {
break;
}
- pos.col += utfc_ptr2len(linep + pos.col);
+ pos.col += utfc_ptr2len((char *)linep + pos.col);
}
if (!findc) {
// no brace in the line, maybe use " #if" then
- if (!cpo_match && *skipwhite(linep) == '#') {
+ if (!cpo_match && *skipwhite((char *)linep) == '#') {
hash_dir = 1;
} else {
return NULL;
@@ -1917,7 +1907,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
oap->motion_type = kMTLineWise; // Linewise for this case only
}
if (initc != '#') {
- ptr = skipwhite(skipwhite(linep) + 1);
+ ptr = (char_u *)skipwhite(skipwhite((char *)linep) + 1);
if (STRNCMP(ptr, "if", 2) == 0 || STRNCMP(ptr, "el", 2) == 0) {
hash_dir = 1;
} else if (STRNCMP(ptr, "endif", 5) == 0) {
@@ -1938,12 +1928,12 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
pos.lnum += hash_dir;
linep = ml_get(pos.lnum);
line_breakcheck(); // check for CTRL-C typed
- ptr = skipwhite(linep);
+ ptr = (char_u *)skipwhite((char *)linep);
if (*ptr != '#') {
continue;
}
pos.col = (colnr_T)(ptr - linep);
- ptr = skipwhite(ptr + 1);
+ ptr = (char_u *)skipwhite((char *)ptr + 1);
if (hash_dir > 0) {
if (STRNCMP(ptr, "if", 2) == 0) {
count++;
@@ -1978,7 +1968,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
// This is just guessing: when 'rightleft' is set, search for a matching
// paren/brace in the other direction.
- if (curwin->w_p_rl && vim_strchr((char_u *)"()[]{}<>", initc) != NULL) {
+ if (curwin->w_p_rl && vim_strchr("()[]{}<>", initc) != NULL) {
backwards = !backwards;
}
@@ -1989,13 +1979,13 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
clearpos(&match_pos);
// backward search: Check if this line contains a single-line comment
- if ((backwards && comment_dir)
- || lisp) {
+ if ((backwards && comment_dir) || lisp) {
comment_col = check_linecomment(linep);
}
if (lisp && comment_col != MAXCOL && pos.col > (colnr_T)comment_col) {
lispcomm = true; // find match inside this comment
}
+
while (!got_int) {
/*
* Go to the next position, forward or backward. We could use
@@ -2022,8 +2012,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
line_breakcheck();
// Check if this line contains a single-line comment
- if (comment_dir
- || lisp) {
+ if (comment_dir || lisp) {
comment_col = check_linecomment(linep);
}
// skip comment
@@ -2037,7 +2026,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
} else { // forward search
if (linep[pos.col] == NUL
// at end of line, go to next one
- // don't search for match in comment
+ // For lisp don't search for match in comment
|| (lisp && comment_col != MAXCOL
&& pos.col == (colnr_T)comment_col)) {
if (pos.lnum == curbuf->b_ml.ml_line_count // end of file
@@ -2060,7 +2049,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
comment_col = check_linecomment(linep);
}
} else {
- pos.col += utfc_ptr2len(linep + pos.col);
+ pos.col += utfc_ptr2len((char *)linep + pos.col);
}
}
@@ -2091,7 +2080,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
} else if (raw_string) {
if (linep[pos.col - 1] == 'R'
&& linep[pos.col] == '"'
- && vim_strchr(linep + pos.col + 1, '(') != NULL) {
+ && vim_strchr((char *)linep + pos.col + 1, '(') != NULL) {
// Possible start of raw string. Now that we have the
// delimiter we can check if it ends before where we
// started searching, or before the previously found
@@ -2205,7 +2194,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
* inquote if the number of quotes in a line is even, unless this
* line or the previous one ends in a '\'. Complicated, isn't it?
*/
- const int c = utf_ptr2char(linep + pos.col);
+ const int c = utf_ptr2char((char *)linep + pos.col);
switch (c) {
case NUL:
// at end of line without trailing backslash, reset inquote
@@ -2272,7 +2261,7 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
* (actually, we skip #\( et al)
*/
if (curbuf->b_p_lisp
- && vim_strchr((char_u *)"(){}[]", c) != NULL
+ && vim_strchr("(){}[]", c) != NULL
&& pos.col > 1
&& check_prevcol(linep, pos.col, '\\', NULL)
&& check_prevcol(linep, pos.col - 1, '#', NULL)) {
@@ -2313,20 +2302,17 @@ pos_T *findmatchlimit(oparg_T *oap, int initc, int flags, int64_t maxtravel)
return (pos_T *)NULL; // never found it
}
-/*
- * Check if line[] contains a / / comment.
- * Return MAXCOL if not, otherwise return the column.
- * TODO: skip strings.
- */
-static int check_linecomment(const char_u *line)
+/// Check if line[] contains a / / comment.
+/// @returns MAXCOL if not, otherwise return the column.
+int check_linecomment(const char_u *line)
{
const char_u *p = line; // scan from start
// skip Lispish one-line comments
if (curbuf->b_p_lisp) {
- if (vim_strchr(p, ';') != NULL) { // there may be comments
+ if (vim_strchr((char *)p, ';') != NULL) { // there may be comments
bool in_str = false; // inside of string
- while ((p = vim_strpbrk(p, (char_u *)"\";")) != NULL) {
+ while ((p = (char_u *)strpbrk((char *)p, "\";")) != NULL) {
if (*p == '"') {
if (in_str) {
if (*(p - 1) != '\\') { // skip escaped quote
@@ -2338,7 +2324,8 @@ static int check_linecomment(const char_u *line)
in_str = true;
}
} else if (!in_str && ((p - line) < 2
- || (*(p - 1) != '\\' && *(p - 2) != '#'))) {
+ || (*(p - 1) != '\\' && *(p - 2) != '#'))
+ && !is_pos_in_string(line, (colnr_T)(p - line))) {
break; // found!
}
p++;
@@ -2347,10 +2334,12 @@ static int check_linecomment(const char_u *line)
p = NULL;
}
} else {
- while ((p = vim_strchr(p, '/')) != NULL) {
- // accept a double /, unless it's preceded with * and followed by *,
- // because * / / * is an end and start of a C comment
- if (p[1] == '/' && (p == line || p[-1] != '*' || p[2] != '*')) {
+ while ((p = (char_u *)vim_strchr((char *)p, '/')) != NULL) {
+ // Accept a double /, unless it's preceded with * and followed by *,
+ // because * / / * is an end and start of a C comment. Only
+ // accept the position if it is not inside a string.
+ if (p[1] == '/' && (p == line || p[-1] != '*' || p[2] != '*')
+ && !is_pos_in_string(line, (colnr_T)(p - line))) {
break;
}
++p;
@@ -2387,14 +2376,14 @@ void showmatch(int c)
*/
// 'matchpairs' is "x:y,x:y"
for (p = curbuf->b_p_mps; *p != NUL; p++) {
- if (utf_ptr2char(p) == c && (curwin->w_p_rl ^ p_ri)) {
+ if (utf_ptr2char((char *)p) == c && (curwin->w_p_rl ^ p_ri)) {
break;
}
- p += utfc_ptr2len(p) + 1;
- if (utf_ptr2char(p) == c && !(curwin->w_p_rl ^ p_ri)) {
+ p += utfc_ptr2len((char *)p) + 1;
+ if (utf_ptr2char((char *)p) == c && !(curwin->w_p_rl ^ p_ri)) {
break;
}
- p += utfc_ptr2len(p);
+ p += utfc_ptr2len((char *)p);
if (*p == NUL) {
return;
}
@@ -2427,7 +2416,7 @@ void showmatch(int c)
save_dollar_vcol = dollar_vcol;
save_state = State;
- State = SHOWMATCH;
+ State = MODE_SHOWMATCH;
ui_cursor_shape(); // may show different cursor shape
curwin->w_cursor = mpos; // move to matching char
*so = 0; // don't use 'scrolloff' here
@@ -2506,7 +2495,7 @@ int findsent(Direction dir, long count)
// go back to the previous non-white non-punctuation character
bool found_dot = false;
while (c = gchar_pos(&pos), ascii_iswhite(c)
- || vim_strchr((char_u *)".!?)]\"'", c) != NULL) {
+ || vim_strchr(".!?)]\"'", c) != NULL) {
tpos = pos;
if (decl(&tpos) == -1 || (LINEEMPTY(tpos.lnum) && dir == FORWARD)) {
break;
@@ -2514,11 +2503,11 @@ int findsent(Direction dir, long count)
if (found_dot) {
break;
}
- if (vim_strchr((char_u *)".!?", c) != NULL) {
+ if (vim_strchr(".!?", c) != NULL) {
found_dot = true;
}
- if (vim_strchr((char_u *)")]\"'", c) != NULL
- && vim_strchr((char_u *)".!?)]\"'", gchar_pos(&tpos)) == NULL) {
+ if (vim_strchr(")]\"'", c) != NULL
+ && vim_strchr(".!?)]\"'", gchar_pos(&tpos)) == NULL) {
break;
}
decl(&pos);
@@ -2542,9 +2531,8 @@ int findsent(Direction dir, long count)
if ((c = inc(&tpos)) == -1) {
break;
}
- }
- while (vim_strchr((char_u *)")]\"'", c = gchar_pos(&tpos))
- != NULL);
+ } while (vim_strchr(")]\"'", c = gchar_pos(&tpos))
+ != NULL);
if (c == -1 || (!cpo_J && (c == ' ' || c == '\t')) || c == NUL
|| (cpo_J && (c == ' ' && inc(&tpos) >= 0
&& gchar_pos(&tpos) == ' '))) {
@@ -3399,7 +3387,7 @@ int current_block(oparg_T *oap, long count, int include, int what, int other)
pos_T start_pos;
pos_T *end_pos;
pos_T old_start, old_end;
- char_u *save_cpo;
+ char *save_cpo;
bool sol = false; // '{' at start of line
old_pos = curwin->w_cursor;
@@ -3434,7 +3422,7 @@ int current_block(oparg_T *oap, long count, int include, int what, int other)
// Ignore quotes here. Keep the "M" flag in 'cpo', as that is what the
// user wants.
save_cpo = p_cpo;
- p_cpo = (char_u *)(vim_strchr(p_cpo, CPO_MATCHBSL) != NULL ? "%M" : "%");
+ p_cpo = vim_strchr(p_cpo, CPO_MATCHBSL) != NULL ? "%M" : "%";
if ((pos = findmatch(NULL, what)) != NULL) {
while (count-- > 0) {
if ((pos = findmatch(NULL, what)) == NULL) {
@@ -3479,11 +3467,11 @@ int current_block(oparg_T *oap, long count, int include, int what, int other)
}
}
- /*
- * In Visual mode, when the resulting area is not bigger than what we
- * started with, extend it to the next block, and then exclude again.
- */
+ // In Visual mode, when the resulting area is not bigger than what we
+ // started with, extend it to the next block, and then exclude again.
+ // Don't try to expand the area if the area is empty.
if (!lt(start_pos, old_start) && !lt(old_end, curwin->w_cursor)
+ && !equalpos(start_pos, curwin->w_cursor)
&& VIsual_active) {
curwin->w_cursor = old_start;
decl(&curwin->w_cursor);
@@ -3533,7 +3521,6 @@ int current_block(oparg_T *oap, long count, int include, int what, int other)
return OK;
}
-
/// @param end_tag when true, return true if the cursor is on "</aaa>".
///
/// @return true if the cursor is on a "<aaa>" tag. Ignore "<aaa/>".
@@ -3672,8 +3659,7 @@ again:
p = get_cursor_pos_ptr();
for (cp = p;
*cp != NUL && *cp != '>' && !ascii_iswhite(*cp);
- MB_PTR_ADV(cp)) {
- }
+ MB_PTR_ADV(cp)) {}
len = (int)(cp - p);
if (len == 0) {
curwin->w_cursor = old_pos;
@@ -3952,7 +3938,6 @@ extend:
return OK;
}
-
/// Search quote char from string line[col].
/// Quote character escaped by one of the characters in "escape" is not counted
/// as a quote.
@@ -3968,12 +3953,15 @@ static int find_next_quote(char_u *line, int col, int quotechar, char_u *escape)
c = line[col];
if (c == NUL) {
return -1;
- } else if (escape != NULL && vim_strchr(escape, c)) {
+ } else if (escape != NULL && vim_strchr((char *)escape, c)) {
col++;
+ if (line[col] == NUL) {
+ return -1;
+ }
} else if (c == quotechar) {
break;
}
- col += utfc_ptr2len(line + col);
+ col += utfc_ptr2len((char *)line + col);
}
return col;
}
@@ -3994,15 +3982,14 @@ static int find_prev_quote(char_u *line, int col_start, int quotechar, char_u *e
col_start -= utf_head_off(line, line + col_start);
n = 0;
if (escape != NULL) {
- while (col_start - n > 0 && vim_strchr(escape,
+ while (col_start - n > 0 && vim_strchr((char *)escape,
line[col_start - n - 1]) != NULL) {
++n;
}
}
if (n & 1) {
col_start -= n; // uneven number of escape chars, skip it
- } else if (line[col_start] ==
- quotechar) {
+ } else if (line[col_start] == quotechar) {
break;
}
}
@@ -4083,6 +4070,11 @@ bool current_quote(oparg_T *oap, long count, bool include, int quotechar)
// Find out if we have a quote in the selection.
while (i <= col_end) {
+ // check for going over the end of the line, which can happen if
+ // the line was changed after the Visual area was selected.
+ if (line[i] == NUL) {
+ break;
+ }
if (line[i++] == quotechar) {
selected_quote = true;
break;
@@ -4120,8 +4112,7 @@ bool current_quote(oparg_T *oap, long count, bool include, int quotechar)
col_end = curwin->w_cursor.col;
}
}
- } else if (line[col_start] == quotechar
- || !vis_empty) {
+ } else if (line[col_start] == quotechar || !vis_empty) {
int first_col = col_start;
if (!vis_empty) {
@@ -4190,9 +4181,8 @@ bool current_quote(oparg_T *oap, long count, bool include, int quotechar)
// Set start position. After vi" another i" must include the ".
// For v2i" include the quotes.
- if (!include && count < 2
- && (vis_empty || !inside_quotes)) {
- ++col_start;
+ if (!include && count < 2 && (vis_empty || !inside_quotes)) {
+ col_start++;
}
curwin->w_cursor.col = col_start;
if (VIsual_active) {
@@ -4268,7 +4258,6 @@ abort_search:
return false;
}
-
/// Find next search match under cursor, cursor at end.
/// Used while an operator is pending, and in Visual mode.
///
@@ -4418,7 +4407,7 @@ static int is_zero_width(char_u *pattern, int move, pos_T *cur, Direction direct
int nmatched = 0;
int result = -1;
pos_T pos;
- int save_called_emsg = called_emsg;
+ const int called_emsg_before = called_emsg;
int flag = 0;
if (pattern == NULL) {
@@ -4444,7 +4433,6 @@ static int is_zero_width(char_u *pattern, int move, pos_T *cur, Direction direct
SEARCH_KEEP + flag, RE_SEARCH, NULL) != FAIL) {
// Zero-width pattern should match somewhere, then we can check if
// start and end are in the same position.
- called_emsg = false;
do {
regmatch.startpos[0].col++;
nmatched = vim_regexec_multi(&regmatch, curwin, curbuf,
@@ -4458,14 +4446,13 @@ static int is_zero_width(char_u *pattern, int move, pos_T *cur, Direction direct
? regmatch.startpos[0].col < pos.col
: regmatch.startpos[0].col > pos.col);
- if (!called_emsg) {
+ if (called_emsg == called_emsg_before) {
result = (nmatched != 0
&& regmatch.startpos[0].lnum == regmatch.endpos[0].lnum
&& regmatch.startpos[0].col == regmatch.endpos[0].col);
}
}
- called_emsg |= save_called_emsg;
vim_regfree(regmatch.regprog);
return result;
}
@@ -4477,7 +4464,7 @@ int linewhite(linenr_T lnum)
{
char_u *p;
- p = skipwhite(ml_get(lnum));
+ p = (char_u *)skipwhite((char *)ml_get(lnum));
return *p == NUL;
}
@@ -4537,7 +4524,7 @@ static void cmdline_search_stat(int dirc, pos_T *pos, pos_T *cursor_pos, bool sh
// keep the message even after redraw, but don't put in history
msg_hist_off = true;
msg_ext_set_kind("search_count");
- give_warning(msgbuf, false);
+ give_warning((char *)msgbuf, false);
msg_hist_off = false;
}
}
@@ -4740,6 +4727,7 @@ void f_searchcount(typval_T *argvars, typval_T *rettv, FunPtr fptr)
}
save_last_search_pattern();
+ save_incsearch_state();
if (pattern != NULL) {
if (*pattern == NUL) {
goto the_end;
@@ -4761,6 +4749,564 @@ void f_searchcount(typval_T *argvars, typval_T *rettv, FunPtr fptr)
the_end:
restore_last_search_pattern();
+ restore_incsearch_state();
+}
+
+/// Fuzzy string matching
+///
+/// Ported from the lib_fts library authored by Forrest Smith.
+/// https://github.com/forrestthewoods/lib_fts/tree/master/code
+///
+/// The following blog describes the fuzzy matching algorithm:
+/// https://www.forrestthewoods.com/blog/reverse_engineering_sublime_texts_fuzzy_match/
+///
+/// Each matching string is assigned a score. The following factors are checked:
+/// - Matched letter
+/// - Unmatched letter
+/// - Consecutively matched letters
+/// - Proximity to start
+/// - Letter following a separator (space, underscore)
+/// - Uppercase letter following lowercase (aka CamelCase)
+///
+/// Matched letters are good. Unmatched letters are bad. Matching near the start
+/// is good. Matching the first letter in the middle of a phrase is good.
+/// Matching the uppercase letters in camel case entries is good.
+///
+/// The score assigned for each factor is explained below.
+/// File paths are different from file names. File extensions may be ignorable.
+/// Single words care about consecutive matches but not separators or camel
+/// case.
+/// Score starts at 100
+/// Matched letter: +0 points
+/// Unmatched letter: -1 point
+/// Consecutive match bonus: +15 points
+/// First letter bonus: +15 points
+/// Separator bonus: +30 points
+/// Camel case bonus: +30 points
+/// Unmatched leading letter: -5 points (max: -15)
+///
+/// There is some nuance to this. Scores don’t have an intrinsic meaning. The
+/// score range isn’t 0 to 100. It’s roughly [50, 150]. Longer words have a
+/// lower minimum score due to unmatched letter penalty. Longer search patterns
+/// have a higher maximum score due to match bonuses.
+///
+/// Separator and camel case bonus is worth a LOT. Consecutive matches are worth
+/// quite a bit.
+///
+/// There is a penalty if you DON’T match the first three letters. Which
+/// effectively rewards matching near the start. However there’s no difference
+/// in matching between the middle and end.
+///
+/// There is not an explicit bonus for an exact match. Unmatched letters receive
+/// a penalty. So shorter strings and closer matches are worth more.
+typedef struct {
+ int idx; ///< used for stable sort
+ listitem_T *item;
+ int score;
+ list_T *lmatchpos;
+} fuzzyItem_T;
+
+/// bonus for adjacent matches; this is higher than SEPARATOR_BONUS so that
+/// matching a whole word is preferred.
+#define SEQUENTIAL_BONUS 40
+/// bonus if match occurs after a path separator
+#define PATH_SEPARATOR_BONUS 30
+/// bonus if match occurs after a word separator
+#define WORD_SEPARATOR_BONUS 25
+/// bonus if match is uppercase and prev is lower
+#define CAMEL_BONUS 30
+/// bonus if the first letter is matched
+#define FIRST_LETTER_BONUS 15
+/// penalty applied for every letter in str before the first match
+#define LEADING_LETTER_PENALTY (-5)
+/// maximum penalty for leading letters
+#define MAX_LEADING_LETTER_PENALTY (-15)
+/// penalty for every letter that doesn't match
+#define UNMATCHED_LETTER_PENALTY (-1)
+/// penalty for gap in matching positions (-2 * k)
+#define GAP_PENALTY (-2)
+/// Score for a string that doesn't fuzzy match the pattern
+#define SCORE_NONE (-9999)
+
+#define FUZZY_MATCH_RECURSION_LIMIT 10
+
+/// Compute a score for a fuzzy matched string. The matching character locations
+/// are in 'matches'.
+static int fuzzy_match_compute_score(const char_u *const str, const int strSz,
+ const uint32_t *const matches, const int numMatches)
+ FUNC_ATTR_NONNULL_ALL FUNC_ATTR_WARN_UNUSED_RESULT FUNC_ATTR_PURE
+{
+ assert(numMatches > 0); // suppress clang "result of operation is garbage"
+ // Initialize score
+ int score = 100;
+
+ // Apply leading letter penalty
+ int penalty = LEADING_LETTER_PENALTY * matches[0];
+ if (penalty < MAX_LEADING_LETTER_PENALTY) {
+ penalty = MAX_LEADING_LETTER_PENALTY;
+ }
+ score += penalty;
+
+ // Apply unmatched penalty
+ const int unmatched = strSz - numMatches;
+ score += UNMATCHED_LETTER_PENALTY * unmatched;
+
+ // Apply ordering bonuses
+ for (int i = 0; i < numMatches; i++) {
+ const uint32_t currIdx = matches[i];
+
+ if (i > 0) {
+ const uint32_t prevIdx = matches[i - 1];
+
+ // Sequential
+ if (currIdx == prevIdx + 1) {
+ score += SEQUENTIAL_BONUS;
+ } else {
+ score += GAP_PENALTY * (currIdx - prevIdx);
+ }
+ }
+
+ // Check for bonuses based on neighbor character value
+ if (currIdx > 0) {
+ // Camel case
+ const char_u *p = str;
+ int neighbor;
+
+ for (uint32_t sidx = 0; sidx < currIdx; sidx++) {
+ neighbor = utf_ptr2char((char *)p);
+ MB_PTR_ADV(p);
+ }
+ const int curr = utf_ptr2char((char *)p);
+
+ if (mb_islower(neighbor) && mb_isupper(curr)) {
+ score += CAMEL_BONUS;
+ }
+
+ // Bonus if the match follows a separator character
+ if (neighbor == '/' || neighbor == '\\') {
+ score += PATH_SEPARATOR_BONUS;
+ } else if (neighbor == ' ' || neighbor == '_') {
+ score += WORD_SEPARATOR_BONUS;
+ }
+ } else {
+ // First letter
+ score += FIRST_LETTER_BONUS;
+ }
+ }
+ return score;
+}
+
+/// Perform a recursive search for fuzzy matching 'fuzpat' in 'str'.
+/// @return the number of matching characters.
+static int fuzzy_match_recursive(const char_u *fuzpat, const char_u *str, uint32_t strIdx,
+ int *const outScore, const char_u *const strBegin,
+ const int strLen, const uint32_t *const srcMatches,
+ uint32_t *const matches, const int maxMatches, int nextMatch,
+ int *const recursionCount)
+ FUNC_ATTR_NONNULL_ARG(1, 2, 4, 5, 8, 11) FUNC_ATTR_WARN_UNUSED_RESULT
+{
+ // Recursion params
+ bool recursiveMatch = false;
+ uint32_t bestRecursiveMatches[MAX_FUZZY_MATCHES];
+ int bestRecursiveScore = 0;
+
+ // Count recursions
+ (*recursionCount)++;
+ if (*recursionCount >= FUZZY_MATCH_RECURSION_LIMIT) {
+ return 0;
+ }
+
+ // Detect end of strings
+ if (*fuzpat == NUL || *str == NUL) {
+ return 0;
+ }
+
+ // Loop through fuzpat and str looking for a match
+ bool first_match = true;
+ while (*fuzpat != NUL && *str != NUL) {
+ const int c1 = utf_ptr2char((char *)fuzpat);
+ const int c2 = utf_ptr2char((char *)str);
+
+ // Found match
+ if (mb_tolower(c1) == mb_tolower(c2)) {
+ // Supplied matches buffer was too short
+ if (nextMatch >= maxMatches) {
+ return 0;
+ }
+
+ // "Copy-on-Write" srcMatches into matches
+ if (first_match && srcMatches != NULL) {
+ memcpy(matches, srcMatches, nextMatch * sizeof(srcMatches[0]));
+ first_match = false;
+ }
+
+ // Recursive call that "skips" this match
+ uint32_t recursiveMatches[MAX_FUZZY_MATCHES];
+ int recursiveScore = 0;
+ const char_u *const next_char = str + utfc_ptr2len((char *)str);
+ if (fuzzy_match_recursive(fuzpat, next_char, strIdx + 1, &recursiveScore, strBegin, strLen,
+ matches, recursiveMatches,
+ sizeof(recursiveMatches) / sizeof(recursiveMatches[0]), nextMatch,
+ recursionCount)) {
+ // Pick best recursive score
+ if (!recursiveMatch || recursiveScore > bestRecursiveScore) {
+ memcpy(bestRecursiveMatches, recursiveMatches,
+ MAX_FUZZY_MATCHES * sizeof(recursiveMatches[0]));
+ bestRecursiveScore = recursiveScore;
+ }
+ recursiveMatch = true;
+ }
+
+ // Advance
+ matches[nextMatch++] = strIdx;
+ MB_PTR_ADV(fuzpat);
+ }
+ MB_PTR_ADV(str);
+ strIdx++;
+ }
+
+ // Determine if full fuzpat was matched
+ const bool matched = *fuzpat == NUL;
+
+ // Calculate score
+ if (matched) {
+ *outScore = fuzzy_match_compute_score(strBegin, strLen, matches, nextMatch);
+ }
+
+ // Return best result
+ if (recursiveMatch && (!matched || bestRecursiveScore > *outScore)) {
+ // Recursive score is better than "this"
+ memcpy(matches, bestRecursiveMatches, maxMatches * sizeof(matches[0]));
+ *outScore = bestRecursiveScore;
+ return nextMatch;
+ } else if (matched) {
+ return nextMatch; // "this" score is better than recursive
+ }
+
+ return 0; // no match
+}
+
+/// fuzzy_match()
+///
+/// Performs exhaustive search via recursion to find all possible matches and
+/// match with highest score.
+/// Scores values have no intrinsic meaning. Possible score range is not
+/// normalized and varies with pattern.
+/// Recursion is limited internally (default=10) to prevent degenerate cases
+/// (pat_arg="aaaaaa" str="aaaaaaaaaaaaaaaaaaaaaaaaaaaaaa").
+/// Uses char_u for match indices. Therefore patterns are limited to
+/// MAX_FUZZY_MATCHES characters.
+///
+/// @return true if 'pat_arg' matches 'str'. Also returns the match score in
+/// 'outScore' and the matching character positions in 'matches'.
+bool fuzzy_match(char_u *const str, const char_u *const pat_arg, const bool matchseq,
+ int *const outScore, uint32_t *const matches, const int maxMatches)
+ FUNC_ATTR_NONNULL_ALL FUNC_ATTR_WARN_UNUSED_RESULT
+{
+ const int len = mb_charlen(str);
+ bool complete = false;
+ int numMatches = 0;
+
+ *outScore = 0;
+
+ char_u *const save_pat = vim_strsave(pat_arg);
+ char_u *pat = save_pat;
+ char_u *p = pat;
+
+ // Try matching each word in 'pat_arg' in 'str'
+ while (true) {
+ if (matchseq) {
+ complete = true;
+ } else {
+ // Extract one word from the pattern (separated by space)
+ p = (char_u *)skipwhite((char *)p);
+ if (*p == NUL) {
+ break;
+ }
+ pat = p;
+ while (*p != NUL && !ascii_iswhite(utf_ptr2char((char *)p))) {
+ MB_PTR_ADV(p);
+ }
+ if (*p == NUL) { // processed all the words
+ complete = true;
+ }
+ *p = NUL;
+ }
+
+ int score = 0;
+ int recursionCount = 0;
+ const int matchCount
+ = fuzzy_match_recursive(pat, str, 0, &score, str, len, NULL, matches + numMatches,
+ maxMatches - numMatches, 0, &recursionCount);
+ if (matchCount == 0) {
+ numMatches = 0;
+ break;
+ }
+
+ // Accumulate the match score and the number of matches
+ *outScore += score;
+ numMatches += matchCount;
+
+ if (complete) {
+ break;
+ }
+
+ // try matching the next word
+ p++;
+ }
+
+ xfree(save_pat);
+ return numMatches != 0;
+}
+
+/// Sort the fuzzy matches in the descending order of the match score.
+/// For items with same score, retain the order using the index (stable sort)
+static int fuzzy_match_item_compare(const void *const s1, const void *const s2)
+ FUNC_ATTR_NONNULL_ALL FUNC_ATTR_WARN_UNUSED_RESULT FUNC_ATTR_PURE
+{
+ const int v1 = ((const fuzzyItem_T *)s1)->score;
+ const int v2 = ((const fuzzyItem_T *)s2)->score;
+ const int idx1 = ((const fuzzyItem_T *)s1)->idx;
+ const int idx2 = ((const fuzzyItem_T *)s2)->idx;
+
+ return v1 == v2 ? (idx1 - idx2) : v1 > v2 ? -1 : 1;
+}
+
+/// Fuzzy search the string 'str' in a list of 'items' and return the matching
+/// strings in 'fmatchlist'.
+/// If 'matchseq' is true, then for multi-word search strings, match all the
+/// words in sequence.
+/// If 'items' is a list of strings, then search for 'str' in the list.
+/// If 'items' is a list of dicts, then either use 'key' to lookup the string
+/// for each item or use 'item_cb' Funcref function to get the string.
+/// If 'retmatchpos' is true, then return a list of positions where 'str'
+/// matches for each item.
+static void fuzzy_match_in_list(list_T *const l, char_u *const str, const bool matchseq,
+ const char_u *const key, Callback *const item_cb,
+ const bool retmatchpos, list_T *const fmatchlist,
+ const long max_matches)
+ FUNC_ATTR_NONNULL_ARG(2, 5, 7)
+{
+ long len = tv_list_len(l);
+ if (len == 0) {
+ return;
+ }
+ if (max_matches > 0 && len > max_matches) {
+ len = max_matches;
+ }
+
+ fuzzyItem_T *const items = xcalloc(len, sizeof(fuzzyItem_T));
+ long match_count = 0;
+ uint32_t matches[MAX_FUZZY_MATCHES];
+
+ // For all the string items in items, get the fuzzy matching score
+ TV_LIST_ITER(l, li, {
+ if (max_matches > 0 && match_count >= max_matches) {
+ break;
+ }
+
+ char_u *itemstr = NULL;
+ typval_T rettv;
+ rettv.v_type = VAR_UNKNOWN;
+ const typval_T *const tv = TV_LIST_ITEM_TV(li);
+ if (tv->v_type == VAR_STRING) { // list of strings
+ itemstr = (char_u *)tv->vval.v_string;
+ } else if (tv->v_type == VAR_DICT && (key != NULL || item_cb->type != kCallbackNone)) {
+ // For a dict, either use the specified key to lookup the string or
+ // use the specified callback function to get the string.
+ if (key != NULL) {
+ itemstr = (char_u *)tv_dict_get_string(tv->vval.v_dict, (const char *)key, false);
+ } else {
+ typval_T argv[2];
+
+ // Invoke the supplied callback (if any) to get the dict item
+ tv->vval.v_dict->dv_refcount++;
+ argv[0].v_type = VAR_DICT;
+ argv[0].vval.v_dict = tv->vval.v_dict;
+ argv[1].v_type = VAR_UNKNOWN;
+ if (callback_call(item_cb, 1, argv, &rettv)) {
+ if (rettv.v_type == VAR_STRING) {
+ itemstr = (char_u *)rettv.vval.v_string;
+ }
+ }
+ tv_dict_unref(tv->vval.v_dict);
+ }
+ }
+
+ int score;
+ if (itemstr != NULL && fuzzy_match(itemstr, str, matchseq, &score, matches,
+ MAX_FUZZY_MATCHES)) {
+ items[match_count].idx = match_count;
+ items[match_count].item = li;
+ items[match_count].score = score;
+
+ // Copy the list of matching positions in itemstr to a list, if
+ // 'retmatchpos' is set.
+ if (retmatchpos) {
+ items[match_count].lmatchpos = tv_list_alloc(kListLenMayKnow);
+ int j = 0;
+ const char_u *p = str;
+ while (*p != NUL) {
+ if (!ascii_iswhite(utf_ptr2char((char *)p)) || matchseq) {
+ tv_list_append_number(items[match_count].lmatchpos, matches[j]);
+ j++;
+ }
+ MB_PTR_ADV(p);
+ }
+ }
+ match_count++;
+ }
+ tv_clear(&rettv);
+ });
+
+ if (match_count > 0) {
+ // Sort the list by the descending order of the match score
+ qsort(items, match_count, sizeof(fuzzyItem_T), fuzzy_match_item_compare);
+
+ // For matchfuzzy(), return a list of matched strings.
+ // ['str1', 'str2', 'str3']
+ // For matchfuzzypos(), return a list with three items.
+ // The first item is a list of matched strings. The second item
+ // is a list of lists where each list item is a list of matched
+ // character positions. The third item is a list of matching scores.
+ // [['str1', 'str2', 'str3'], [[1, 3], [1, 3], [1, 3]]]
+ list_T *retlist;
+ if (retmatchpos) {
+ const listitem_T *const li = tv_list_find(fmatchlist, 0);
+ assert(li != NULL && TV_LIST_ITEM_TV(li)->vval.v_list != NULL);
+ retlist = TV_LIST_ITEM_TV(li)->vval.v_list;
+ } else {
+ retlist = fmatchlist;
+ }
+
+ // Copy the matching strings with a valid score to the return list
+ for (long i = 0; i < match_count; i++) {
+ if (items[i].score == SCORE_NONE) {
+ break;
+ }
+ tv_list_append_tv(retlist, TV_LIST_ITEM_TV(items[i].item));
+ }
+
+ // next copy the list of matching positions
+ if (retmatchpos) {
+ const listitem_T *li = tv_list_find(fmatchlist, -2);
+ assert(li != NULL && TV_LIST_ITEM_TV(li)->vval.v_list != NULL);
+ retlist = TV_LIST_ITEM_TV(li)->vval.v_list;
+
+ for (long i = 0; i < match_count; i++) {
+ if (items[i].score == SCORE_NONE) {
+ break;
+ }
+ tv_list_append_list(retlist, items[i].lmatchpos);
+ }
+
+ // copy the matching scores
+ li = tv_list_find(fmatchlist, -1);
+ assert(li != NULL && TV_LIST_ITEM_TV(li)->vval.v_list != NULL);
+ retlist = TV_LIST_ITEM_TV(li)->vval.v_list;
+ for (long i = 0; i < match_count; i++) {
+ if (items[i].score == SCORE_NONE) {
+ break;
+ }
+ tv_list_append_number(retlist, items[i].score);
+ }
+ }
+ }
+ xfree(items);
+}
+
+/// Do fuzzy matching. Returns the list of matched strings in 'rettv'.
+/// If 'retmatchpos' is true, also returns the matching character positions.
+static void do_fuzzymatch(const typval_T *const argvars, typval_T *const rettv,
+ const bool retmatchpos)
+ FUNC_ATTR_NONNULL_ALL
+{
+ // validate and get the arguments
+ if (argvars[0].v_type != VAR_LIST || argvars[0].vval.v_list == NULL) {
+ semsg(_(e_listarg), retmatchpos ? "matchfuzzypos()" : "matchfuzzy()");
+ return;
+ }
+ if (argvars[1].v_type != VAR_STRING || argvars[1].vval.v_string == NULL) {
+ semsg(_(e_invarg2), tv_get_string(&argvars[1]));
+ return;
+ }
+
+ Callback cb = CALLBACK_NONE;
+ const char_u *key = NULL;
+ bool matchseq = false;
+ long max_matches = 0;
+ if (argvars[2].v_type != VAR_UNKNOWN) {
+ if (argvars[2].v_type != VAR_DICT || argvars[2].vval.v_dict == NULL) {
+ emsg(_(e_dictreq));
+ return;
+ }
+
+ // To search a dict, either a callback function or a key can be
+ // specified.
+ dict_T *const d = argvars[2].vval.v_dict;
+ const dictitem_T *di;
+ if ((di = tv_dict_find(d, "key", -1)) != NULL) {
+ if (di->di_tv.v_type != VAR_STRING || di->di_tv.vval.v_string == NULL
+ || *di->di_tv.vval.v_string == NUL) {
+ semsg(_(e_invarg2), tv_get_string(&di->di_tv));
+ return;
+ }
+ key = (const char_u *)tv_get_string(&di->di_tv);
+ } else if (!tv_dict_get_callback(d, "text_cb", -1, &cb)) {
+ semsg(_(e_invargval), "text_cb");
+ return;
+ }
+
+ if ((di = tv_dict_find(d, "limit", -1)) != NULL) {
+ if (di->di_tv.v_type != VAR_NUMBER) {
+ semsg(_(e_invarg2), tv_get_string(&di->di_tv));
+ return;
+ }
+ max_matches = (long)tv_get_number_chk(&di->di_tv, NULL);
+ }
+
+ if (tv_dict_find(d, "matchseq", -1) != NULL) {
+ matchseq = true;
+ }
+ }
+
+ // get the fuzzy matches
+ tv_list_alloc_ret(rettv, retmatchpos ? 3 : kListLenUnknown);
+ if (retmatchpos) {
+ // For matchfuzzypos(), a list with three items are returned. First
+ // item is a list of matching strings, the second item is a list of
+ // lists with matching positions within each string and the third item
+ // is the list of scores of the matches.
+ tv_list_append_list(rettv->vval.v_list, tv_list_alloc(kListLenUnknown));
+ tv_list_append_list(rettv->vval.v_list, tv_list_alloc(kListLenUnknown));
+ tv_list_append_list(rettv->vval.v_list, tv_list_alloc(kListLenUnknown));
+ }
+
+ fuzzy_match_in_list(argvars[0].vval.v_list, (char_u *)tv_get_string(&argvars[1]), matchseq, key,
+ &cb, retmatchpos, rettv->vval.v_list, max_matches);
+ callback_free(&cb);
+}
+
+/// "matchfuzzy()" function
+void f_matchfuzzy(typval_T *argvars, typval_T *rettv, FunPtr fptr)
+{
+ do_fuzzymatch(argvars, rettv, false);
+}
+
+/// "matchfuzzypos()" function
+void f_matchfuzzypos(typval_T *argvars, typval_T *rettv, FunPtr fptr)
+{
+ do_fuzzymatch(argvars, rettv, true);
+}
+
+/// Get line "lnum" and copy it into "buf[LSIZE]".
+/// The copy is made because the regexp may make the line invalid when using a
+/// mark.
+static char_u *get_line_and_copy(linenr_T lnum, char_u *buf)
+{
+ char_u *line = ml_get(lnum);
+ STRLCPY(buf, line, LSIZE);
+ return buf;
}
/// Find identifiers or defines in included files.
@@ -4785,7 +5331,7 @@ void find_pattern_in_path(char_u *ptr, Direction dir, size_t len, bool whole, bo
char_u *pat;
char_u *new_fname;
- char_u *curr_fname = curbuf->b_fname;
+ char_u *curr_fname = (char_u *)curbuf->b_fname;
char_u *prev_fname = NULL;
linenr_T lnum;
int depth;
@@ -4825,7 +5371,7 @@ void find_pattern_in_path(char_u *ptr, Direction dir, size_t len, bool whole, bo
sprintf((char *)pat, whole ? "\\<%.*s\\>" : "%.*s", (int)len, ptr);
// ignore case according to p_ic, p_scs and pat
regmatch.rm_ic = ignorecase(pat);
- regmatch.regprog = vim_regcomp(pat, p_magic ? RE_MAGIC : 0);
+ regmatch.regprog = vim_regcomp((char *)pat, p_magic ? RE_MAGIC : 0);
xfree(pat);
if (regmatch.regprog == NULL) {
goto fpip_end;
@@ -4833,7 +5379,7 @@ void find_pattern_in_path(char_u *ptr, Direction dir, size_t len, bool whole, bo
}
inc_opt = (*curbuf->b_p_inc == NUL) ? p_inc : curbuf->b_p_inc;
if (*inc_opt != NUL) {
- incl_regmatch.regprog = vim_regcomp(inc_opt, p_magic ? RE_MAGIC : 0);
+ incl_regmatch.regprog = vim_regcomp((char *)inc_opt, p_magic ? RE_MAGIC : 0);
if (incl_regmatch.regprog == NULL) {
goto fpip_end;
}
@@ -4841,7 +5387,8 @@ void find_pattern_in_path(char_u *ptr, Direction dir, size_t len, bool whole, bo
}
if (type == FIND_DEFINE && (*curbuf->b_p_def != NUL || *p_def != NUL)) {
def_regmatch.regprog = vim_regcomp(*curbuf->b_p_def == NUL
- ? p_def : curbuf->b_p_def, p_magic ? RE_MAGIC : 0);
+ ? (char *)p_def : (char *)curbuf->b_p_def,
+ p_magic ? RE_MAGIC : 0);
if (def_regmatch.regprog == NULL) {
goto fpip_end;
}
@@ -4858,13 +5405,13 @@ void find_pattern_in_path(char_u *ptr, Direction dir, size_t len, bool whole, bo
if (lnum > end_lnum) { // do at least one line
lnum = end_lnum;
}
- line = ml_get(lnum);
+ line = get_line_and_copy(lnum, file_line);
for (;;) {
if (incl_regmatch.regprog != NULL
- && vim_regexec(&incl_regmatch, line, (colnr_T)0)) {
- char_u *p_fname = (curr_fname == curbuf->b_fname)
- ? curbuf->b_ffname : curr_fname;
+ && vim_regexec(&incl_regmatch, (char *)line, (colnr_T)0)) {
+ char_u *p_fname = (curr_fname == (char_u *)curbuf->b_fname)
+ ? (char_u *)curbuf->b_ffname : curr_fname;
if (inc_opt != NULL && strstr((char *)inc_opt, "\\zs") != NULL) {
// Use text from '\zs' to '\ze' (or end) of 'include'.
@@ -4888,8 +5435,8 @@ void find_pattern_in_path(char_u *ptr, Direction dir, size_t len, bool whole, bo
if (i == max_path_depth) {
break;
}
- if (path_full_compare(new_fname, files[i].name,
- true, true) & kEqualFiles) {
+ if (path_full_compare((char *)new_fname, (char *)files[i].name, true,
+ true) & kEqualFiles) {
if (type != CHECK_PATH
&& action == ACTION_SHOW_ALL && files[i].matched) {
msg_putchar('\n'); // cursor below last one */
@@ -4951,10 +5498,8 @@ void find_pattern_in_path(char_u *ptr, Direction dir, size_t len, bool whole, bo
} else {
// find the file name after the end of the match
for (p = incl_regmatch.endp[0];
- *p && !vim_isfilec(*p); p++) {
- }
- for (i = 0; vim_isfilec(p[i]); i++) {
- }
+ *p && !vim_isfilec(*p); p++) {}
+ for (i = 0; vim_isfilec(p[i]); i++) {}
}
if (i == 0) {
@@ -5044,12 +5589,10 @@ void find_pattern_in_path(char_u *ptr, Direction dir, size_t len, bool whole, bo
search_line:
define_matched = false;
if (def_regmatch.regprog != NULL
- && vim_regexec(&def_regmatch, line, (colnr_T)0)) {
- /*
- * Pattern must be first identifier after 'define', so skip
- * to that position before checking for match of pattern. Also
- * don't let it match beyond the end of this identifier.
- */
+ && vim_regexec(&def_regmatch, (char *)line, (colnr_T)0)) {
+ // Pattern must be first identifier after 'define', so skip
+ // to that position before checking for match of pattern. Also
+ // don't let it match beyond the end of this identifier.
p = def_regmatch.endp[0];
while (*p && !vim_iswordc(*p)) {
p++;
@@ -5065,7 +5608,7 @@ search_line:
if (define_matched
|| (compl_cont_status & CONT_SOL)) {
// compare the first "len" chars from "ptr"
- startp = skipwhite(p);
+ startp = (char_u *)skipwhite((char *)p);
if (p_ic) {
matched = !mb_strnicmp(startp, ptr, len);
} else {
@@ -5076,7 +5619,7 @@ search_line:
matched = false;
}
} else if (regmatch.regprog != NULL
- && vim_regexec(&regmatch, line, (colnr_T)(p - line))) {
+ && vim_regexec(&regmatch, (char *)line, (colnr_T)(p - line))) {
matched = true;
startp = regmatch.startp[0];
// Check if the line is not a comment line (unless we are
@@ -5084,8 +5627,8 @@ search_line:
// is not considered to be a comment line.
if (skip_comments) {
if ((*line != '#'
- || STRNCMP(skipwhite(line + 1), "define", 6) != 0)
- && get_leader_len(line, NULL, false, true)) {
+ || STRNCMP(skipwhite((char *)line + 1), "define", 6) != 0)
+ && get_leader_len((char *)line, NULL, false, true)) {
matched = false;
}
@@ -5095,7 +5638,7 @@ search_line:
* * /" when looking for "normal".
* Note: Doesn't skip "/ *" in comments.
*/
- p = skipwhite(line);
+ p = (char_u *)skipwhite((char *)line);
if (matched
|| (p[0] == '/' && p[1] == '*') || p[0] == '*') {
for (p = line; *p && p < startp; ++p) {
@@ -5150,7 +5693,7 @@ search_line:
if (lnum >= end_lnum) {
goto exit_matched;
}
- line = ml_get(++lnum);
+ line = get_line_and_copy(++lnum, file_line);
} else if (vim_fgets(line = file_line,
LSIZE, files[depth].fp)) {
goto exit_matched;
@@ -5158,20 +5701,20 @@ search_line:
// we read a line, set "already" to check this "line" later
// if depth >= 0 we'll increase files[depth].lnum far
- // bellow -- Acevedo
- already = aux = p = skipwhite(line);
+ // below -- Acevedo
+ already = aux = p = (char_u *)skipwhite((char *)line);
p = find_word_start(p);
p = find_word_end(p);
if (p > aux) {
- if (*aux != ')' && IObuff[i-1] != TAB) {
- if (IObuff[i-1] != ' ') {
+ if (*aux != ')' && IObuff[i - 1] != TAB) {
+ if (IObuff[i - 1] != ' ') {
IObuff[i++] = ' ';
}
// IObuf =~ "\(\k\|\i\).* ", thus i >= 2
if (p_js
- && (IObuff[i-2] == '.'
- || IObuff[i-2] == '?'
- || IObuff[i-2] == '!')) {
+ && (IObuff[i - 2] == '.'
+ || IObuff[i - 2] == '?'
+ || IObuff[i - 2] == '!')) {
IObuff[i++] = ' ';
}
}
@@ -5192,7 +5735,8 @@ search_line:
}
const int add_r = ins_compl_add_infercase(aux, i, p_ic,
- curr_fname == curbuf->b_fname ? NULL : curr_fname,
+ curr_fname == (char_u *)curbuf->b_fname
+ ? NULL : curr_fname,
dir, cont_s_ipos);
if (add_r == OK) {
// if dir was BACKWARD then honor it just once
@@ -5266,7 +5810,7 @@ search_line:
curwin->w_cursor.lnum = lnum;
check_cursor();
} else {
- if (!GETFILE_SUCCESS(getfile(0, files[depth].name, NULL, true,
+ if (!GETFILE_SUCCESS(getfile(0, (char *)files[depth].name, NULL, true,
files[depth].lnum, false))) {
break; // failed to jump to file
}
@@ -5297,7 +5841,7 @@ exit_matched:
&& action == ACTION_EXPAND
&& !(compl_cont_status & CONT_SOL)
&& *startp != NUL
- && *(p = startp + utfc_ptr2len(startp)) != NUL) {
+ && *(p = startp + utfc_ptr2len((char *)startp)) != NUL) {
goto search_line;
}
}
@@ -5320,8 +5864,8 @@ exit_matched:
--old_files;
files[old_files].name = files[depth].name;
files[old_files].matched = files[depth].matched;
- --depth;
- curr_fname = (depth == -1) ? curbuf->b_fname
+ depth--;
+ curr_fname = (depth == -1) ? (char_u *)curbuf->b_fname
: files[depth].name;
if (depth < depth_displayed) {
depth_displayed = depth;
@@ -5341,7 +5885,7 @@ exit_matched:
if (++lnum > end_lnum) {
break;
}
- line = ml_get(lnum);
+ line = get_line_and_copy(lnum, file_line);
}
already = NULL;
}
@@ -5415,7 +5959,7 @@ static void show_pat_in_path(char_u *line, int type, bool did_show, int action,
if (action == ACTION_SHOW_ALL) {
snprintf((char *)IObuff, IOSIZE, "%3ld: ", count); // Show match nr.
msg_puts((const char *)IObuff);
- snprintf((char *)IObuff, IOSIZE, "%4ld", *lnum); // Show line nr.
+ snprintf((char *)IObuff, IOSIZE, "%4" PRIdLINENR, *lnum); // Show line nr.
// Highlight line numbers.
msg_puts_attr((const char *)IObuff, HL_ATTR(HLF_N));
msg_puts(" ");
@@ -5486,3 +6030,9 @@ bool search_was_last_used(void)
{
return last_idx == 0;
}
+
+/// @return true if 'hlsearch' highlight is currently in use.
+bool using_hlsearch(void)
+{
+ return spats[last_idx].pat != NULL && p_hls && !no_hlsearch;
+}
generated by cgit (git 2.34.1) at 2025-08-10 21:17:42 +0000