X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;ds=sidebyside;f=src%2Flanguage%2Flexer%2Fsegment.c;h=9c2c9b11e8eba18e34b13f40124e3eb3439a97ae;hb=76e0b3df102f54672f8d4dfcc19c144e52a8ff4e;hp=8d17ce38fe56b4de73d05b764c652a9d9a040594;hpb=012b9e4ecf47d844aea352b2c0b7dda80a7194bd;p=pspp diff --git a/src/language/lexer/segment.c b/src/language/lexer/segment.c index 8d17ce38fe..9c2c9b11e8 100644 --- a/src/language/lexer/segment.c +++ b/src/language/lexer/segment.c @@ -38,7 +38,9 @@ enum segmenter_state S_DOCUMENT_1, S_DOCUMENT_2, S_DOCUMENT_3, - S_FILE_LABEL, + S_FILE_LABEL_1, + S_FILE_LABEL_2, + S_FILE_LABEL_3, S_DO_REPEAT_1, S_DO_REPEAT_2, S_DO_REPEAT_3, @@ -50,8 +52,6 @@ enum segmenter_state S_BEGIN_DATA_2, S_BEGIN_DATA_3, S_BEGIN_DATA_4, - S_TITLE_1, - S_TITLE_2 }; #define SS_START_OF_LINE (1u << 0) @@ -227,7 +227,7 @@ is_all_spaces (const char *input_, size_t n) for (int ofs = 0; ofs < n; ofs += mblen) { ucs4_t uc; - mblen = u8_mbtouc (&uc, input, n); + mblen = u8_mbtouc (&uc, input + ofs, n - ofs); if (!lex_uc_is_space (uc)) return false; } @@ -759,18 +759,6 @@ segmenter_parse_id__ (struct segmenter *s, const char *input, size_t n, *type = SEG_START_DOCUMENT; return 0; } - else if (lex_id_match (ss_cstr ("TITLE"), word) - || lex_id_match (ss_cstr ("SUBTITLE"), word)) - { - int result = segmenter_unquoted (input, n, eof, ofs); - if (result < 0) - return -1; - else if (result) - { - s->state = S_TITLE_1; - return ofs; - } - } else if (lex_id_match_n (ss_cstr ("DEFINE"), word, 6)) { s->state = S_DEFINE_1; @@ -784,7 +772,7 @@ segmenter_parse_id__ (struct segmenter *s, const char *input, size_t n, return -1; else if (lex_id_match (ss_cstr ("LABEL"), ss_cstr (id))) { - s->state = S_FILE_LABEL; + s->state = S_FILE_LABEL_1; s->substate = 0; return ofs; } @@ -1257,9 +1245,9 @@ segmenter_parse_start_of_line__ (struct segmenter *s, } static int -segmenter_parse_file_label__ (struct segmenter *s, - const char *input, size_t n, bool eof, - enum segment_type *type) +segmenter_parse_file_label_1__ (struct segmenter *s, + const char *input, size_t n, bool eof, + enum segment_type *type) { struct segmenter sub; int ofs; @@ -1282,7 +1270,7 @@ segmenter_parse_file_label__ (struct segmenter *s, else { if (result) - s->state = S_TITLE_1; + s->state = S_FILE_LABEL_2; else *s = sub; return ofs; @@ -1295,6 +1283,70 @@ segmenter_parse_file_label__ (struct segmenter *s, } } +static int +segmenter_parse_file_label_2__ (struct segmenter *s, + const char *input, size_t n, bool eof, + enum segment_type *type) +{ + int ofs; + + ofs = skip_spaces (input, n, eof, 0); + if (ofs < 0) + return -1; + s->state = S_FILE_LABEL_3; + *type = SEG_SPACES; + return ofs; +} + +static int +segmenter_parse_file_label_3__ (struct segmenter *s, + const char *input, size_t n, bool eof, + enum segment_type *type) +{ + int endcmd; + int ofs; + + endcmd = -1; + ofs = 0; + while (ofs < n) + { + ucs4_t uc; + int mblen; + + mblen = segmenter_u8_to_uc__ (&uc, input, n, eof, ofs); + if (mblen < 0) + return -1; + + switch (uc) + { + case '\n': + goto end_of_line; + + case '.': + endcmd = ofs; + break; + + default: + if (!lex_uc_is_space (uc)) + endcmd = -1; + break; + } + + ofs += mblen; + } + + if (eof) + { + end_of_line: + s->state = S_GENERAL; + s->substate = 0; + *type = SEG_UNQUOTED_STRING; + return endcmd >= 0 ? endcmd : ofs; + } + + return -1; +} + static int segmenter_subparse (struct segmenter *s, const char *input, size_t n, bool eof, @@ -1512,7 +1564,10 @@ segmenter_parse_define_2__ (struct segmenter *s, { s->nest--; if (!s->nest) - s->state = S_DEFINE_3; + { + s->state = S_DEFINE_3; + s->substate = 0; + } return ofs; } @@ -1524,11 +1579,33 @@ find_enddefine (struct substring input) { size_t n = input.length; const struct substring enddefine = ss_cstr ("!ENDDEFINE"); - for (size_t i = 0; i + enddefine.length <= n; i++) - if (input.string[i] == '!' - && ss_equals_case (ss_substr (input, i, enddefine.length), enddefine)) - return i; - return SIZE_MAX; + for (int ofs = 0;;) + { + /* Skip !ENDDEFINE in comment. */ + ofs = skip_spaces_and_comments (input.string, n, true, ofs); + if (ofs + enddefine.length > n) + return SIZE_MAX; + + char c = input.string[ofs]; + if (c == '!' + && ss_equals_case (ss_substr (input, ofs, enddefine.length), + enddefine)) + return ofs; + else if (c == '\'' || c == '"') + { + /* Skip quoted !ENDDEFINE. */ + ofs++; + for (;;) + { + if (ofs >= n) + return SIZE_MAX; + else if (input.string[ofs++] == c) + break; + } + } + else + ofs++; + } } /* We are in the body of a macro definition, looking for additional lines of @@ -1554,9 +1631,15 @@ segmenter_parse_define_3__ (struct segmenter *s, The line might be blank, whether completely empty or just spaces and comments. That's OK: we need to report blank lines because they can - have significance. */ - *type = SEG_MACRO_BODY; + have significance. + + However, if the first line of the macro body (the same line as the + closing parenthesis in the argument definition) is blank, we just + report it as spaces because it's not significant. */ + *type = (s->substate == 0 && is_all_spaces (input, ofs) + ? SEG_SPACES : SEG_MACRO_BODY); s->state = S_DEFINE_4; + s->substate = 1; return ofs; } else @@ -1708,70 +1791,6 @@ segmenter_parse_begin_data_4__ (struct segmenter *s, return ofs; } -static int -segmenter_parse_title_1__ (struct segmenter *s, - const char *input, size_t n, bool eof, - enum segment_type *type) -{ - int ofs; - - ofs = skip_spaces (input, n, eof, 0); - if (ofs < 0) - return -1; - s->state = S_TITLE_2; - *type = SEG_SPACES; - return ofs; -} - -static int -segmenter_parse_title_2__ (struct segmenter *s, - const char *input, size_t n, bool eof, - enum segment_type *type) -{ - int endcmd; - int ofs; - - endcmd = -1; - ofs = 0; - while (ofs < n) - { - ucs4_t uc; - int mblen; - - mblen = segmenter_u8_to_uc__ (&uc, input, n, eof, ofs); - if (mblen < 0) - return -1; - - switch (uc) - { - case '\n': - goto end_of_line; - - case '.': - endcmd = ofs; - break; - - default: - if (!lex_uc_is_space (uc)) - endcmd = -1; - break; - } - - ofs += mblen; - } - - if (eof) - { - end_of_line: - s->state = S_GENERAL; - s->substate = 0; - *type = SEG_UNQUOTED_STRING; - return endcmd >= 0 ? endcmd : ofs; - } - - return -1; -} - /* Returns the name of segment TYPE as a string. The caller must not modify or free the returned string. @@ -1789,17 +1808,28 @@ segment_type_to_string (enum segment_type type) } } -/* Initializes S as a segmenter with the given syntax MODE. +/* Returns a segmenter with the given syntax MODE. + + If IS_SNIPPET is false, then the segmenter will parse as if it's being given + a whole file. This means, for example, that it will interpret - or + at the + beginning of the syntax as a separator between commands (since - or + at the + beginning of a line has this meaning). + + If IS_SNIPPET is true, then the segmenter will parse as if it's being given + an isolated piece of syntax. This means that, for example, that it will + interpret - or + at the beginning of the syntax as an operator token or (if + followed by a digit) as part of a number. A segmenter does not contain any external references, so nothing needs to be done to destroy one. For the same reason, segmenters may be copied with plain struct assignment (or memcpy). */ -void -segmenter_init (struct segmenter *s, enum segmenter_mode mode) +struct segmenter +segmenter_init (enum segmenter_mode mode, bool is_snippet) { - s->state = S_SHBANG; - s->substate = 0; - s->mode = mode; + return (struct segmenter) { + .state = is_snippet ? S_GENERAL : S_SHBANG, + .mode = mode, + }; } /* Returns the mode passed to segmenter_init() for S. */ @@ -1872,8 +1902,12 @@ segmenter_push (struct segmenter *s, const char *input, size_t n, bool eof, case S_DOCUMENT_3: return segmenter_parse_document_3__ (s, type); - case S_FILE_LABEL: - return segmenter_parse_file_label__ (s, input, n, eof, type); + case S_FILE_LABEL_1: + return segmenter_parse_file_label_1__ (s, input, n, eof, type); + case S_FILE_LABEL_2: + return segmenter_parse_file_label_2__ (s, input, n, eof, type); + case S_FILE_LABEL_3: + return segmenter_parse_file_label_3__ (s, input, n, eof, type); case S_DO_REPEAT_1: return segmenter_parse_do_repeat_1__ (s, input, n, eof, type); @@ -1899,11 +1933,6 @@ segmenter_push (struct segmenter *s, const char *input, size_t n, bool eof, return segmenter_parse_begin_data_3__ (s, input, n, eof, type); case S_BEGIN_DATA_4: return segmenter_parse_begin_data_4__ (s, input, n, eof, type); - - case S_TITLE_1: - return segmenter_parse_title_1__ (s, input, n, eof, type); - case S_TITLE_2: - return segmenter_parse_title_2__ (s, input, n, eof, type); } NOT_REACHED (); @@ -1934,8 +1963,11 @@ segmenter_get_prompt (const struct segmenter *s) case S_DOCUMENT_3: return PROMPT_FIRST; - case S_FILE_LABEL: + case S_FILE_LABEL_1: return PROMPT_LATER; + case S_FILE_LABEL_2: + case S_FILE_LABEL_3: + return PROMPT_FIRST; case S_DO_REPEAT_1: case S_DO_REPEAT_2: @@ -1958,9 +1990,6 @@ segmenter_get_prompt (const struct segmenter *s) case S_BEGIN_DATA_4: return PROMPT_DATA; - case S_TITLE_1: - case S_TITLE_2: - return PROMPT_FIRST; } NOT_REACHED ();