diff options
Diffstat (limited to 'src/nvim/syntax.c')
| -rw-r--r-- | src/nvim/syntax.c | 7818 |
1 files changed, 7818 insertions, 0 deletions
diff --git a/src/nvim/syntax.c b/src/nvim/syntax.c new file mode 100644 index 0000000000..41128868fc --- /dev/null +++ b/src/nvim/syntax.c @@ -0,0 +1,7818 @@ +/* + * VIM - Vi IMproved by Bram Moolenaar + * + * Do ":help uganda" in Vim to read copying and usage conditions. + * Do ":help credits" in Vim to see a list of people who contributed. + * See README.txt for an overview of the Vim source code. + */ + +/* + * syntax.c: code for syntax highlighting + */ + +#include <string.h> +#include <stdlib.h> + +#include "vim.h" +#include "syntax.h" +#include "charset.h" +#include "eval.h" +#include "ex_cmds2.h" +#include "ex_docmd.h" +#include "fileio.h" +#include "fold.h" +#include "hashtab.h" +#include "indent_c.h" +#include "mbyte.h" +#include "memline.h" +#include "memory.h" +#include "message.h" +#include "misc1.h" +#include "misc2.h" +#include "keymap.h" +#include "garray.h" +#include "option.h" +#include "os_unix.h" +#include "path.h" +#include "regexp.h" +#include "screen.h" +#include "term.h" +#include "ui.h" +#include "os/os.h" + +/* + * Structure that stores information about a highlight group. + * The ID of a highlight group is also called group ID. It is the index in + * the highlight_ga array PLUS ONE. + */ +struct hl_group { + char_u *sg_name; /* highlight group name */ + char_u *sg_name_u; /* uppercase of sg_name */ + /* for normal terminals */ + int sg_term; /* "term=" highlighting attributes */ + char_u *sg_start; /* terminal string for start highl */ + char_u *sg_stop; /* terminal string for stop highl */ + int sg_term_attr; /* Screen attr for term mode */ + /* for color terminals */ + int sg_cterm; /* "cterm=" highlighting attr */ + int sg_cterm_bold; /* bold attr was set for light color */ + int sg_cterm_fg; /* terminal fg color number + 1 */ + int sg_cterm_bg; /* terminal bg color number + 1 */ + int sg_cterm_attr; /* Screen attr for color term mode */ + /* Store the sp color name for the GUI or synIDattr() */ + int sg_gui; /* "gui=" highlighting attributes */ + char_u *sg_gui_fg_name; /* GUI foreground color name */ + char_u *sg_gui_bg_name; /* GUI background color name */ + char_u *sg_gui_sp_name; /* GUI special color name */ + int sg_link; /* link to this highlight group ID */ + int sg_set; /* combination of SG_* flags */ + scid_T sg_scriptID; /* script in which the group was last set */ +}; + +#define SG_TERM 1 /* term has been set */ +#define SG_CTERM 2 /* cterm has been set */ +#define SG_GUI 4 /* gui has been set */ +#define SG_LINK 8 /* link has been set */ + +static garray_T highlight_ga; /* highlight groups for 'highlight' option */ + +#define HL_TABLE() ((struct hl_group *)((highlight_ga.ga_data))) + +#define MAX_HL_ID 20000 /* maximum value for a highlight ID. */ + +/* Flags to indicate an additional string for highlight name completion. */ +static int include_none = 0; /* when 1 include "None" */ +static int include_default = 0; /* when 1 include "default" */ +static int include_link = 0; /* when 2 include "link" and "clear" */ + +/* + * The "term", "cterm" and "gui" arguments can be any combination of the + * following names, separated by commas (but no spaces!). + */ +static char *(hl_name_table[]) = +{"bold", "standout", "underline", "undercurl", + "italic", "reverse", "inverse", "NONE"}; +static int hl_attr_table[] = +{HL_BOLD, HL_STANDOUT, HL_UNDERLINE, HL_UNDERCURL, HL_ITALIC, HL_INVERSE, + HL_INVERSE, 0}; + +static int get_attr_entry(garray_T *table, attrentry_T *aep); +static void syn_unadd_group(void); +static void set_hl_attr(int idx); +static void highlight_list_one(int id); +static int highlight_list_arg(int id, int didh, int type, int iarg, + char_u *sarg, + char *name); +static int syn_add_group(char_u *name); +static int syn_list_header(int did_header, int outlen, int id); +static int hl_has_settings(int idx, int check_link); +static void highlight_clear(int idx); + + +/* + * An attribute number is the index in attr_table plus ATTR_OFF. + */ +#define ATTR_OFF (HL_ALL + 1) + + +#define SYN_NAMELEN 50 /* maximum length of a syntax name */ + +/* different types of offsets that are possible */ +#define SPO_MS_OFF 0 /* match start offset */ +#define SPO_ME_OFF 1 /* match end offset */ +#define SPO_HS_OFF 2 /* highl. start offset */ +#define SPO_HE_OFF 3 /* highl. end offset */ +#define SPO_RS_OFF 4 /* region start offset */ +#define SPO_RE_OFF 5 /* region end offset */ +#define SPO_LC_OFF 6 /* leading context offset */ +#define SPO_COUNT 7 + +static char *(spo_name_tab[SPO_COUNT]) = +{"ms=", "me=", "hs=", "he=", "rs=", "re=", "lc="}; + +/* + * The patterns that are being searched for are stored in a syn_pattern. + * A match item consists of one pattern. + * A start/end item consists of n start patterns and m end patterns. + * A start/skip/end item consists of n start patterns, one skip pattern and m + * end patterns. + * For the latter two, the patterns are always consecutive: start-skip-end. + * + * A character offset can be given for the matched text (_m_start and _m_end) + * and for the actually highlighted text (_h_start and _h_end). + */ +typedef struct syn_pattern { + char sp_type; /* see SPTYPE_ defines below */ + char sp_syncing; /* this item used for syncing */ + int sp_flags; /* see HL_ defines below */ + int sp_cchar; /* conceal substitute character */ + struct sp_syn sp_syn; /* struct passed to in_id_list() */ + short sp_syn_match_id; /* highlight group ID of pattern */ + char_u *sp_pattern; /* regexp to match, pattern */ + regprog_T *sp_prog; /* regexp to match, program */ + syn_time_T sp_time; + int sp_ic; /* ignore-case flag for sp_prog */ + short sp_off_flags; /* see below */ + int sp_offsets[SPO_COUNT]; /* offsets */ + short *sp_cont_list; /* cont. group IDs, if non-zero */ + short *sp_next_list; /* next group IDs, if non-zero */ + int sp_sync_idx; /* sync item index (syncing only) */ + int sp_line_id; /* ID of last line where tried */ + int sp_startcol; /* next match in sp_line_id line */ +} synpat_T; + +/* The sp_off_flags are computed like this: + * offset from the start of the matched text: (1 << SPO_XX_OFF) + * offset from the end of the matched text: (1 << (SPO_XX_OFF + SPO_COUNT)) + * When both are present, only one is used. + */ + +#define SPTYPE_MATCH 1 /* match keyword with this group ID */ +#define SPTYPE_START 2 /* match a regexp, start of item */ +#define SPTYPE_END 3 /* match a regexp, end of item */ +#define SPTYPE_SKIP 4 /* match a regexp, skip within item */ + + +#define SYN_ITEMS(buf) ((synpat_T *)((buf)->b_syn_patterns.ga_data)) + +#define NONE_IDX -2 /* value of sp_sync_idx for "NONE" */ + +/* + * Flags for b_syn_sync_flags: + */ +#define SF_CCOMMENT 0x01 /* sync on a C-style comment */ +#define SF_MATCH 0x02 /* sync by matching a pattern */ + +#define SYN_STATE_P(ssp) ((bufstate_T *)((ssp)->ga_data)) + +#define MAXKEYWLEN 80 /* maximum length of a keyword */ + +/* + * The attributes of the syntax item that has been recognized. + */ +static int current_attr = 0; /* attr of current syntax word */ +static int current_id = 0; /* ID of current char for syn_get_id() */ +static int current_trans_id = 0; /* idem, transparency removed */ +static int current_flags = 0; +static int current_seqnr = 0; +static int current_sub_char = 0; + +typedef struct syn_cluster_S { + char_u *scl_name; /* syntax cluster name */ + char_u *scl_name_u; /* uppercase of scl_name */ + short *scl_list; /* IDs in this syntax cluster */ +} syn_cluster_T; + +/* + * Methods of combining two clusters + */ +#define CLUSTER_REPLACE 1 /* replace first list with second */ +#define CLUSTER_ADD 2 /* add second list to first */ +#define CLUSTER_SUBTRACT 3 /* subtract second list from first */ + +#define SYN_CLSTR(buf) ((syn_cluster_T *)((buf)->b_syn_clusters.ga_data)) + +/* + * Syntax group IDs have different types: + * 0 - 19999 normal syntax groups + * 20000 - 20999 ALLBUT indicator (current_syn_inc_tag added) + * 21000 - 21999 TOP indicator (current_syn_inc_tag added) + * 22000 - 22999 CONTAINED indicator (current_syn_inc_tag added) + * 23000 - 32767 cluster IDs (subtract SYNID_CLUSTER for the cluster ID) + */ +#define SYNID_ALLBUT MAX_HL_ID /* syntax group ID for contains=ALLBUT */ +#define SYNID_TOP 21000 /* syntax group ID for contains=TOP */ +#define SYNID_CONTAINED 22000 /* syntax group ID for contains=CONTAINED */ +#define SYNID_CLUSTER 23000 /* first syntax group ID for clusters */ + +#define MAX_SYN_INC_TAG 999 /* maximum before the above overflow */ +#define MAX_CLUSTER_ID (32767 - SYNID_CLUSTER) + +/* + * Annoying Hack(TM): ":syn include" needs this pointer to pass to + * expand_filename(). Most of the other syntax commands don't need it, so + * instead of passing it to them, we stow it here. + */ +static char_u **syn_cmdlinep; + +/* + * Another Annoying Hack(TM): To prevent rules from other ":syn include"'d + * files from leaking into ALLBUT lists, we assign a unique ID to the + * rules in each ":syn include"'d file. + */ +static int current_syn_inc_tag = 0; +static int running_syn_inc_tag = 0; + +/* + * In a hashtable item "hi_key" points to "keyword" in a keyentry. + * This avoids adding a pointer to the hashtable item. + * KE2HIKEY() converts a var pointer to a hashitem key pointer. + * HIKEY2KE() converts a hashitem key pointer to a var pointer. + * HI2KE() converts a hashitem pointer to a var pointer. + */ +static keyentry_T dumkey; +#define KE2HIKEY(kp) ((kp)->keyword) +#define HIKEY2KE(p) ((keyentry_T *)((p) - (dumkey.keyword - (char_u *)&dumkey))) +#define HI2KE(hi) HIKEY2KE((hi)->hi_key) + +/* + * To reduce the time spent in keepend(), remember at which level in the state + * stack the first item with "keepend" is present. When "-1", there is no + * "keepend" on the stack. + */ +static int keepend_level = -1; + +static char msg_no_items[] = N_("No Syntax items defined for this buffer"); + +/* + * For the current state we need to remember more than just the idx. + * When si_m_endpos.lnum is 0, the items other than si_idx are unknown. + * (The end positions have the column number of the next char) + */ +typedef struct state_item { + int si_idx; /* index of syntax pattern or + KEYWORD_IDX */ + int si_id; /* highlight group ID for keywords */ + int si_trans_id; /* idem, transparency removed */ + int si_m_lnum; /* lnum of the match */ + int si_m_startcol; /* starting column of the match */ + lpos_T si_m_endpos; /* just after end posn of the match */ + lpos_T si_h_startpos; /* start position of the highlighting */ + lpos_T si_h_endpos; /* end position of the highlighting */ + lpos_T si_eoe_pos; /* end position of end pattern */ + int si_end_idx; /* group ID for end pattern or zero */ + int si_ends; /* if match ends before si_m_endpos */ + int si_attr; /* attributes in this state */ + long si_flags; /* HL_HAS_EOL flag in this state, and + * HL_SKIP* for si_next_list */ + int si_seqnr; /* sequence number */ + int si_cchar; /* substitution character for conceal */ + short *si_cont_list; /* list of contained groups */ + short *si_next_list; /* nextgroup IDs after this item ends */ + reg_extmatch_T *si_extmatch; /* \z(...\) matches from start + * pattern */ +} stateitem_T; + +#define KEYWORD_IDX -1 /* value of si_idx for keywords */ +#define ID_LIST_ALL (short *)-1 /* valid of si_cont_list for containing all + but contained groups */ + +static int next_seqnr = 0; /* value to use for si_seqnr */ + +/* + * Struct to reduce the number of arguments to get_syn_options(), it's used + * very often. + */ +typedef struct { + int flags; /* flags for contained and transparent */ + int keyword; /* TRUE for ":syn keyword" */ + int *sync_idx; /* syntax item for "grouphere" argument, NULL + if not allowed */ + char has_cont_list; /* TRUE if "cont_list" can be used */ + short *cont_list; /* group IDs for "contains" argument */ + short *cont_in_list; /* group IDs for "containedin" argument */ + short *next_list; /* group IDs for "nextgroup" argument */ +} syn_opt_arg_T; + +/* + * The next possible match in the current line for any pattern is remembered, + * to avoid having to try for a match in each column. + * If next_match_idx == -1, not tried (in this line) yet. + * If next_match_col == MAXCOL, no match found in this line. + * (All end positions have the column of the char after the end) + */ +static int next_match_col; /* column for start of next match */ +static lpos_T next_match_m_endpos; /* position for end of next match */ +static lpos_T next_match_h_startpos; /* pos. for highl. start of next match */ +static lpos_T next_match_h_endpos; /* pos. for highl. end of next match */ +static int next_match_idx; /* index of matched item */ +static long next_match_flags; /* flags for next match */ +static lpos_T next_match_eos_pos; /* end of start pattn (start region) */ +static lpos_T next_match_eoe_pos; /* pos. for end of end pattern */ +static int next_match_end_idx; /* ID of group for end pattn or zero */ +static reg_extmatch_T *next_match_extmatch = NULL; + +/* + * A state stack is an array of integers or stateitem_T, stored in a + * garray_T. A state stack is invalid if it's itemsize entry is zero. + */ +#define INVALID_STATE(ssp) ((ssp)->ga_itemsize == 0) +#define VALID_STATE(ssp) ((ssp)->ga_itemsize != 0) + +/* + * The current state (within the line) of the recognition engine. + * When current_state.ga_itemsize is 0 the current state is invalid. + */ +static win_T *syn_win; /* current window for highlighting */ +static buf_T *syn_buf; /* current buffer for highlighting */ +static synblock_T *syn_block; /* current buffer for highlighting */ +static linenr_T current_lnum = 0; /* lnum of current state */ +static colnr_T current_col = 0; /* column of current state */ +static int current_state_stored = 0; /* TRUE if stored current state + * after setting current_finished */ +static int current_finished = 0; /* current line has been finished */ +static garray_T current_state /* current stack of state_items */ + = {0, 0, 0, 0, NULL}; +static short *current_next_list = NULL; /* when non-zero, nextgroup list */ +static int current_next_flags = 0; /* flags for current_next_list */ +static int current_line_id = 0; /* unique number for current line */ + +#define CUR_STATE(idx) ((stateitem_T *)(current_state.ga_data))[idx] + +static void syn_sync(win_T *wp, linenr_T lnum, synstate_T *last_valid); +static int syn_match_linecont(linenr_T lnum); +static void syn_start_line(void); +static void syn_update_ends(int startofline); +static void syn_stack_alloc(void); +static int syn_stack_cleanup(void); +static void syn_stack_free_entry(synblock_T *block, synstate_T *p); +static synstate_T *syn_stack_find_entry(linenr_T lnum); +static synstate_T *store_current_state(void); +static void load_current_state(synstate_T *from); +static void invalidate_current_state(void); +static int syn_stack_equal(synstate_T *sp); +static void validate_current_state(void); +static int syn_finish_line(int syncing); +static int syn_current_attr(int syncing, int displaying, int *can_spell, + int keep_state); +static int did_match_already(int idx, garray_T *gap); +static stateitem_T *push_next_match(stateitem_T *cur_si); +static void check_state_ends(void); +static void update_si_attr(int idx); +static void check_keepend(void); +static void update_si_end(stateitem_T *sip, int startcol, int force); +static short *copy_id_list(short *list); +static int in_id_list(stateitem_T *item, short *cont_list, + struct sp_syn *ssp, + int contained); +static void push_current_state(int idx); +static void pop_current_state(void); +static void syn_clear_time(syn_time_T *tt); +static void syntime_clear(void); +static int syn_compare_syntime(const void *v1, const void *v2); +static void syntime_report(void); +static int syn_time_on = FALSE; +# define IF_SYN_TIME(p) (p) + +static void syn_stack_apply_changes_block(synblock_T *block, buf_T *buf); +static void find_endpos(int idx, lpos_T *startpos, lpos_T *m_endpos, + lpos_T *hl_endpos, long *flagsp, lpos_T *end_endpos, + int *end_idx, reg_extmatch_T *start_ext); +static void clear_syn_state(synstate_T *p); +static void clear_current_state(void); + +static void limit_pos(lpos_T *pos, lpos_T *limit); +static void limit_pos_zero(lpos_T *pos, lpos_T *limit); +static void syn_add_end_off(lpos_T *result, regmmatch_T *regmatch, + synpat_T *spp, int idx, + int extra); +static void syn_add_start_off(lpos_T *result, regmmatch_T *regmatch, + synpat_T *spp, int idx, + int extra); +static char_u *syn_getcurline(void); +static int syn_regexec(regmmatch_T *rmp, linenr_T lnum, colnr_T col, + syn_time_T *st); +static int check_keyword_id(char_u *line, int startcol, int *endcol, + long *flags, short **next_list, + stateitem_T *cur_si, + int *ccharp); +static keyentry_T *match_keyword(char_u *keyword, hashtab_T *ht, + stateitem_T *cur_si); +static void syn_cmd_case(exarg_T *eap, int syncing); +static void syn_cmd_spell(exarg_T *eap, int syncing); +static void syntax_sync_clear(void); +static void syn_remove_pattern(synblock_T *block, int idx); +static void syn_clear_pattern(synblock_T *block, int i); +static void syn_clear_cluster(synblock_T *block, int i); +static void syn_cmd_clear(exarg_T *eap, int syncing); +static void syn_cmd_conceal(exarg_T *eap, int syncing); +static void syn_clear_one(int id, int syncing); +static void syn_cmd_on(exarg_T *eap, int syncing); +static void syn_cmd_enable(exarg_T *eap, int syncing); +static void syn_cmd_reset(exarg_T *eap, int syncing); +static void syn_cmd_manual(exarg_T *eap, int syncing); +static void syn_cmd_off(exarg_T *eap, int syncing); +static void syn_cmd_onoff(exarg_T *eap, char *name); +static void syn_cmd_list(exarg_T *eap, int syncing); +static void syn_lines_msg(void); +static void syn_match_msg(void); +static void syn_stack_free_block(synblock_T *block); +static void syn_list_one(int id, int syncing, int link_only); +static void syn_list_cluster(int id); +static void put_id_list(char_u *name, short *list, int attr); +static void put_pattern(char *s, int c, synpat_T *spp, int attr); +static int syn_list_keywords(int id, hashtab_T *ht, int did_header, + int attr); +static void syn_clear_keyword(int id, hashtab_T *ht); +static void clear_keywtab(hashtab_T *ht); +static void add_keyword(char_u *name, int id, int flags, + short *cont_in_list, short *next_list, + int conceal_char); +static char_u *get_group_name(char_u *arg, char_u **name_end); +static char_u *get_syn_options(char_u *arg, syn_opt_arg_T *opt, + int *conceal_char); +static void syn_cmd_include(exarg_T *eap, int syncing); +static void syn_cmd_keyword(exarg_T *eap, int syncing); +static void syn_cmd_match(exarg_T *eap, int syncing); +static void syn_cmd_region(exarg_T *eap, int syncing); +static int syn_compare_stub(const void *v1, const void *v2); +static void syn_cmd_cluster(exarg_T *eap, int syncing); +static int syn_scl_name2id(char_u *name); +static int syn_scl_namen2id(char_u *linep, int len); +static int syn_check_cluster(char_u *pp, int len); +static int syn_add_cluster(char_u *name); +static void init_syn_patterns(void); +static char_u *get_syn_pattern(char_u *arg, synpat_T *ci); +static void syn_cmd_sync(exarg_T *eap, int syncing); +static int get_id_list(char_u **arg, int keylen, short **list); +static void syn_combine_list(short **clstr1, short **clstr2, + int list_op); +static void syn_incl_toplevel(int id, int *flagsp); + +/* + * Start the syntax recognition for a line. This function is normally called + * from the screen updating, once for each displayed line. + * The buffer is remembered in syn_buf, because get_syntax_attr() doesn't get + * it. Careful: curbuf and curwin are likely to point to another buffer and + * window. + */ +void syntax_start(win_T *wp, linenr_T lnum) +{ + synstate_T *p; + synstate_T *last_valid = NULL; + synstate_T *last_min_valid = NULL; + synstate_T *sp, *prev = NULL; + linenr_T parsed_lnum; + linenr_T first_stored; + int dist; + static int changedtick = 0; /* remember the last change ID */ + + current_sub_char = NUL; + + /* + * After switching buffers, invalidate current_state. + * Also do this when a change was made, the current state may be invalid + * then. + */ + if (syn_block != wp->w_s || changedtick != syn_buf->b_changedtick) { + invalidate_current_state(); + syn_buf = wp->w_buffer; + syn_block = wp->w_s; + } + changedtick = syn_buf->b_changedtick; + syn_win = wp; + + /* + * Allocate syntax stack when needed. + */ + syn_stack_alloc(); + if (syn_block->b_sst_array == NULL) + return; /* out of memory */ + syn_block->b_sst_lasttick = display_tick; + + /* + * If the state of the end of the previous line is useful, store it. + */ + if (VALID_STATE(¤t_state) + && current_lnum < lnum + && current_lnum < syn_buf->b_ml.ml_line_count) { + (void)syn_finish_line(FALSE); + if (!current_state_stored) { + ++current_lnum; + (void)store_current_state(); + } + + /* + * If the current_lnum is now the same as "lnum", keep the current + * state (this happens very often!). Otherwise invalidate + * current_state and figure it out below. + */ + if (current_lnum != lnum) + invalidate_current_state(); + } else + invalidate_current_state(); + + /* + * Try to synchronize from a saved state in b_sst_array[]. + * Only do this if lnum is not before and not to far beyond a saved state. + */ + if (INVALID_STATE(¤t_state) && syn_block->b_sst_array != NULL) { + /* Find last valid saved state before start_lnum. */ + for (p = syn_block->b_sst_first; p != NULL; p = p->sst_next) { + if (p->sst_lnum > lnum) + break; + if (p->sst_lnum <= lnum && p->sst_change_lnum == 0) { + last_valid = p; + if (p->sst_lnum >= lnum - syn_block->b_syn_sync_minlines) + last_min_valid = p; + } + } + if (last_min_valid != NULL) + load_current_state(last_min_valid); + } + + /* + * If "lnum" is before or far beyond a line with a saved state, need to + * re-synchronize. + */ + if (INVALID_STATE(¤t_state)) { + syn_sync(wp, lnum, last_valid); + if (current_lnum == 1) + /* First line is always valid, no matter "minlines". */ + first_stored = 1; + else + /* Need to parse "minlines" lines before state can be considered + * valid to store. */ + first_stored = current_lnum + syn_block->b_syn_sync_minlines; + } else + first_stored = current_lnum; + + /* + * Advance from the sync point or saved state until the current line. + * Save some entries for syncing with later on. + */ + if (syn_block->b_sst_len <= Rows) + dist = 999999; + else + dist = syn_buf->b_ml.ml_line_count / (syn_block->b_sst_len - Rows) + 1; + while (current_lnum < lnum) { + syn_start_line(); + (void)syn_finish_line(FALSE); + ++current_lnum; + + /* If we parsed at least "minlines" lines or started at a valid + * state, the current state is considered valid. */ + if (current_lnum >= first_stored) { + /* Check if the saved state entry is for the current line and is + * equal to the current state. If so, then validate all saved + * states that depended on a change before the parsed line. */ + if (prev == NULL) + prev = syn_stack_find_entry(current_lnum - 1); + if (prev == NULL) + sp = syn_block->b_sst_first; + else + sp = prev; + while (sp != NULL && sp->sst_lnum < current_lnum) + sp = sp->sst_next; + if (sp != NULL + && sp->sst_lnum == current_lnum + && syn_stack_equal(sp)) { + parsed_lnum = current_lnum; + prev = sp; + while (sp != NULL && sp->sst_change_lnum <= parsed_lnum) { + if (sp->sst_lnum <= lnum) + /* valid state before desired line, use this one */ + prev = sp; + else if (sp->sst_change_lnum == 0) + /* past saved states depending on change, break here. */ + break; + sp->sst_change_lnum = 0; + sp = sp->sst_next; + } + load_current_state(prev); + } + /* Store the state at this line when it's the first one, the line + * where we start parsing, or some distance from the previously + * saved state. But only when parsed at least 'minlines'. */ + else if (prev == NULL + || current_lnum == lnum + || current_lnum >= prev->sst_lnum + dist) + prev = store_current_state(); + } + + /* This can take a long time: break when CTRL-C pressed. The current + * state will be wrong then. */ + line_breakcheck(); + if (got_int) { + current_lnum = lnum; + break; + } + } + + syn_start_line(); +} + +/* + * We cannot simply discard growarrays full of state_items or buf_states; we + * have to manually release their extmatch pointers first. + */ +static void clear_syn_state(synstate_T *p) +{ + int i; + garray_T *gap; + + if (p->sst_stacksize > SST_FIX_STATES) { + gap = &(p->sst_union.sst_ga); + for (i = 0; i < gap->ga_len; i++) + unref_extmatch(SYN_STATE_P(gap)[i].bs_extmatch); + ga_clear(gap); + } else { + for (i = 0; i < p->sst_stacksize; i++) + unref_extmatch(p->sst_union.sst_stack[i].bs_extmatch); + } +} + +/* + * Cleanup the current_state stack. + */ +static void clear_current_state(void) +{ + int i; + stateitem_T *sip; + + sip = (stateitem_T *)(current_state.ga_data); + for (i = 0; i < current_state.ga_len; i++) + unref_extmatch(sip[i].si_extmatch); + ga_clear(¤t_state); +} + +/* + * Try to find a synchronisation point for line "lnum". + * + * This sets current_lnum and the current state. One of three methods is + * used: + * 1. Search backwards for the end of a C-comment. + * 2. Search backwards for given sync patterns. + * 3. Simply start on a given number of lines above "lnum". + */ +static void syn_sync(win_T *wp, linenr_T start_lnum, synstate_T *last_valid) +{ + buf_T *curbuf_save; + win_T *curwin_save; + pos_T cursor_save; + int idx; + linenr_T lnum; + linenr_T end_lnum; + linenr_T break_lnum; + int had_sync_point; + stateitem_T *cur_si; + synpat_T *spp; + char_u *line; + int found_flags = 0; + int found_match_idx = 0; + linenr_T found_current_lnum = 0; + int found_current_col= 0; + lpos_T found_m_endpos; + colnr_T prev_current_col; + + /* + * Clear any current state that might be hanging around. + */ + invalidate_current_state(); + + /* + * Start at least "minlines" back. Default starting point for parsing is + * there. + * Start further back, to avoid that scrolling backwards will result in + * resyncing for every line. Now it resyncs only one out of N lines, + * where N is minlines * 1.5, or minlines * 2 if minlines is small. + * Watch out for overflow when minlines is MAXLNUM. + */ + if (syn_block->b_syn_sync_minlines > start_lnum) + start_lnum = 1; + else { + if (syn_block->b_syn_sync_minlines == 1) + lnum = 1; + else if (syn_block->b_syn_sync_minlines < 10) + lnum = syn_block->b_syn_sync_minlines * 2; + else + lnum = syn_block->b_syn_sync_minlines * 3 / 2; + if (syn_block->b_syn_sync_maxlines != 0 + && lnum > syn_block->b_syn_sync_maxlines) + lnum = syn_block->b_syn_sync_maxlines; + if (lnum >= start_lnum) + start_lnum = 1; + else + start_lnum -= lnum; + } + current_lnum = start_lnum; + + /* + * 1. Search backwards for the end of a C-style comment. + */ + if (syn_block->b_syn_sync_flags & SF_CCOMMENT) { + /* Need to make syn_buf the current buffer for a moment, to be able to + * use find_start_comment(). */ + curwin_save = curwin; + curwin = wp; + curbuf_save = curbuf; + curbuf = syn_buf; + + /* + * Skip lines that end in a backslash. + */ + for (; start_lnum > 1; --start_lnum) { + line = ml_get(start_lnum - 1); + if (*line == NUL || *(line + STRLEN(line) - 1) != '\\') + break; + } + current_lnum = start_lnum; + + /* set cursor to start of search */ + cursor_save = wp->w_cursor; + wp->w_cursor.lnum = start_lnum; + wp->w_cursor.col = 0; + + /* + * If the line is inside a comment, need to find the syntax item that + * defines the comment. + * Restrict the search for the end of a comment to b_syn_sync_maxlines. + */ + if (find_start_comment((int)syn_block->b_syn_sync_maxlines) != NULL) { + for (idx = syn_block->b_syn_patterns.ga_len; --idx >= 0; ) + if (SYN_ITEMS(syn_block)[idx].sp_syn.id + == syn_block->b_syn_sync_id + && SYN_ITEMS(syn_block)[idx].sp_type == SPTYPE_START) { + validate_current_state(); + push_current_state(idx); + update_si_attr(current_state.ga_len - 1); + break; + } + } + + /* restore cursor and buffer */ + wp->w_cursor = cursor_save; + curwin = curwin_save; + curbuf = curbuf_save; + } + /* + * 2. Search backwards for given sync patterns. + */ + else if (syn_block->b_syn_sync_flags & SF_MATCH) { + if (syn_block->b_syn_sync_maxlines != 0 + && start_lnum > syn_block->b_syn_sync_maxlines) + break_lnum = start_lnum - syn_block->b_syn_sync_maxlines; + else + break_lnum = 0; + + found_m_endpos.lnum = 0; + found_m_endpos.col = 0; + end_lnum = start_lnum; + lnum = start_lnum; + while (--lnum > break_lnum) { + /* This can take a long time: break when CTRL-C pressed. */ + line_breakcheck(); + if (got_int) { + invalidate_current_state(); + current_lnum = start_lnum; + break; + } + + /* Check if we have run into a valid saved state stack now. */ + if (last_valid != NULL && lnum == last_valid->sst_lnum) { + load_current_state(last_valid); + break; + } + + /* + * Check if the previous line has the line-continuation pattern. + */ + if (lnum > 1 && syn_match_linecont(lnum - 1)) + continue; + + /* + * Start with nothing on the state stack + */ + validate_current_state(); + + for (current_lnum = lnum; current_lnum < end_lnum; ++current_lnum) { + syn_start_line(); + for (;; ) { + had_sync_point = syn_finish_line(TRUE); + /* + * When a sync point has been found, remember where, and + * continue to look for another one, further on in the line. + */ + if (had_sync_point && current_state.ga_len) { + cur_si = &CUR_STATE(current_state.ga_len - 1); + if (cur_si->si_m_endpos.lnum > start_lnum) { + /* ignore match that goes to after where started */ + current_lnum = end_lnum; + break; + } + if (cur_si->si_idx < 0) { + /* Cannot happen? */ + found_flags = 0; + found_match_idx = KEYWORD_IDX; + } else { + spp = &(SYN_ITEMS(syn_block)[cur_si->si_idx]); + found_flags = spp->sp_flags; + found_match_idx = spp->sp_sync_idx; + } + found_current_lnum = current_lnum; + found_current_col = current_col; + found_m_endpos = cur_si->si_m_endpos; + /* + * Continue after the match (be aware of a zero-length + * match). + */ + if (found_m_endpos.lnum > current_lnum) { + current_lnum = found_m_endpos.lnum; + current_col = found_m_endpos.col; + if (current_lnum >= end_lnum) + break; + } else if (found_m_endpos.col > current_col) + current_col = found_m_endpos.col; + else + ++current_col; + + /* syn_current_attr() will have skipped the check for + * an item that ends here, need to do that now. Be + * careful not to go past the NUL. */ + prev_current_col = current_col; + if (syn_getcurline()[current_col] != NUL) + ++current_col; + check_state_ends(); + current_col = prev_current_col; + } else + break; + } + } + + /* + * If a sync point was encountered, break here. + */ + if (found_flags) { + /* + * Put the item that was specified by the sync point on the + * state stack. If there was no item specified, make the + * state stack empty. + */ + clear_current_state(); + if (found_match_idx >= 0) { + push_current_state(found_match_idx); + update_si_attr(current_state.ga_len - 1); + } + + /* + * When using "grouphere", continue from the sync point + * match, until the end of the line. Parsing starts at + * the next line. + * For "groupthere" the parsing starts at start_lnum. + */ + if (found_flags & HL_SYNC_HERE) { + if (current_state.ga_len) { + cur_si = &CUR_STATE(current_state.ga_len - 1); + cur_si->si_h_startpos.lnum = found_current_lnum; + cur_si->si_h_startpos.col = found_current_col; + update_si_end(cur_si, (int)current_col, TRUE); + check_keepend(); + } + current_col = found_m_endpos.col; + current_lnum = found_m_endpos.lnum; + (void)syn_finish_line(FALSE); + ++current_lnum; + } else + current_lnum = start_lnum; + + break; + } + + end_lnum = lnum; + invalidate_current_state(); + } + + /* Ran into start of the file or exceeded maximum number of lines */ + if (lnum <= break_lnum) { + invalidate_current_state(); + current_lnum = break_lnum + 1; + } + } + + validate_current_state(); +} + +/* + * Return TRUE if the line-continuation pattern matches in line "lnum". + */ +static int syn_match_linecont(linenr_T lnum) +{ + regmmatch_T regmatch; + + if (syn_block->b_syn_linecont_prog != NULL) { + regmatch.rmm_ic = syn_block->b_syn_linecont_ic; + regmatch.regprog = syn_block->b_syn_linecont_prog; + return syn_regexec(®match, lnum, (colnr_T)0, + IF_SYN_TIME(&syn_block->b_syn_linecont_time)); + } + return FALSE; +} + +/* + * Prepare the current state for the start of a line. + */ +static void syn_start_line(void) +{ + current_finished = FALSE; + current_col = 0; + + /* + * Need to update the end of a start/skip/end that continues from the + * previous line and regions that have "keepend". + */ + if (current_state.ga_len > 0) { + syn_update_ends(TRUE); + check_state_ends(); + } + + next_match_idx = -1; + ++current_line_id; +} + +/* + * Check for items in the stack that need their end updated. + * When "startofline" is TRUE the last item is always updated. + * When "startofline" is FALSE the item with "keepend" is forcefully updated. + */ +static void syn_update_ends(int startofline) +{ + stateitem_T *cur_si; + int i; + int seen_keepend;< |