=================================================================== RCS file: /cvs/mandoc/mdoc_markdown.c,v retrieving revision 1.5 retrieving revision 1.18 diff -u -p -r1.5 -r1.18 --- mandoc/mdoc_markdown.c 2017/03/07 13:09:27 1.5 +++ mandoc/mdoc_markdown.c 2017/05/04 17:48:29 1.18 @@ -1,4 +1,4 @@ -/* $Id: mdoc_markdown.c,v 1.5 2017/03/07 13:09:27 schwarze Exp $ */ +/* $Id: mdoc_markdown.c,v 1.18 2017/05/04 17:48:29 schwarze Exp $ */ /* * Copyright (c) 2017 Ingo Schwarze * @@ -19,6 +19,7 @@ #include #include #include +#include #include #include "mandoc_aux.h" @@ -43,6 +44,7 @@ static void md_rawword(const char *); static void md_word(const char *); static void md_named(const char *); static void md_char(unsigned char); +static void md_uri(const char *); static int md_cond_head(struct roff_node *); static int md_cond_body(struct roff_node *); @@ -67,6 +69,7 @@ static int md_pre_Fo(struct roff_node *); static int md_pre_In(struct roff_node *); static int md_pre_It(struct roff_node *); static int md_pre_Lk(struct roff_node *); +static int md_pre_Mt(struct roff_node *); static int md_pre_Nd(struct roff_node *); static int md_pre_Nm(struct roff_node *); static int md_pre_No(struct roff_node *); @@ -90,6 +93,7 @@ static void md_post_En(struct roff_node *); static void md_post_Eo(struct roff_node *); static void md_post_Fa(struct roff_node *); static void md_post_Fd(struct roff_node *); +static void md_post_Fl(struct roff_node *); static void md_post_Fn(struct roff_node *); static void md_post_Fo(struct roff_node *); static void md_post_In(struct roff_node *); @@ -100,8 +104,7 @@ static void md_post_Pf(struct roff_node *); static void md_post_Vt(struct roff_node *); static void md_post__T(struct roff_node *); -static const struct md_act md_acts[MDOC_MAX + 1] = { - { NULL, md_pre_Ap, NULL, NULL, NULL }, /* Ap */ +static const struct md_act __md_acts[MDOC_MAX - MDOC_Dd] = { { NULL, NULL, NULL, NULL, NULL }, /* Dd */ { NULL, NULL, NULL, NULL, NULL }, /* Dt */ { NULL, NULL, NULL, NULL, NULL }, /* Os */ @@ -117,6 +120,7 @@ static const struct md_act md_acts[MDOC_MAX + 1] = { { NULL, md_pre_It, md_post_It, NULL, NULL }, /* It */ { NULL, md_pre_raw, md_post_raw, "*", "*" }, /* Ad */ { NULL, md_pre_An, NULL, NULL, NULL }, /* An */ + { NULL, md_pre_Ap, NULL, NULL, NULL }, /* Ap */ { NULL, md_pre_raw, md_post_raw, "*", "*" }, /* Ar */ { NULL, md_pre_raw, md_post_raw, "**", "**" }, /* Cd */ { NULL, md_pre_raw, md_post_raw, "**", "**" }, /* Cm */ @@ -126,7 +130,7 @@ static const struct md_act md_acts[MDOC_MAX + 1] = { { NULL, NULL, NULL, NULL, NULL }, /* Ex */ { NULL, md_pre_Fa, md_post_Fa, NULL, NULL }, /* Fa */ { NULL, md_pre_Fd, md_post_Fd, "**", "**" }, /* Fd */ - { NULL, md_pre_raw, md_post_raw, "**-", "**" }, /* Fl */ + { NULL, md_pre_raw, md_post_Fl, "**-", "**" }, /* Fl */ { NULL, md_pre_Fn, md_post_Fn, NULL, NULL }, /* Fn */ { NULL, md_pre_Fd, md_post_raw, "*", "*" }, /* Ft */ { NULL, md_pre_raw, md_post_raw, "**", "**" }, /* Ic */ @@ -210,7 +214,7 @@ static const struct md_act md_acts[MDOC_MAX + 1] = { { NULL, NULL, md_post_Lb, NULL, NULL }, /* Lb */ { NULL, md_pre_Pp, NULL, NULL, NULL }, /* Lp */ { NULL, md_pre_Lk, NULL, NULL, NULL }, /* Lk */ - { NULL, md_pre_raw, md_post_raw, "<", ">" }, /* Mt */ + { NULL, md_pre_Mt, NULL, NULL, NULL }, /* Mt */ { md_cond_body, md_pre_word, md_post_word, "{", "}" }, /* Brq */ { md_cond_body, md_pre_word, md_post_word, "{", "}" }, /* Bro */ { NULL, NULL, NULL, NULL, NULL }, /* Brc */ @@ -219,13 +223,12 @@ static const struct md_act md_acts[MDOC_MAX + 1] = { { md_cond_body, md_pre_En, md_post_En, NULL, NULL }, /* En */ { NULL, NULL, NULL, NULL, NULL }, /* Dx */ { NULL, NULL, md_post_pc, NULL, NULL }, /* %Q */ - { NULL, md_pre_br, NULL, NULL, NULL }, /* br */ { NULL, md_pre_Pp, NULL, NULL, NULL }, /* sp */ { NULL, md_pre_Lk, md_post_pc, NULL, NULL }, /* %U */ { NULL, NULL, NULL, NULL, NULL }, /* Ta */ { NULL, NULL, NULL, NULL, NULL }, /* ll */ - { NULL, NULL, NULL, NULL, NULL }, /* ROOT */ }; +static const struct md_act *const md_acts = __md_acts - MDOC_Dd; static int outflags; #define MD_spc (1 << 0) /* Blank character before next word. */ @@ -243,9 +246,9 @@ static int escflags; /* Escape in generated markdown #define ESC_BOL (1 << 0) /* "#*+-" near the beginning of a line. */ #define ESC_NUM (1 << 1) /* "." after a leading number. */ #define ESC_HYP (1 << 2) /* "(" immediately after "]". */ -#define ESC_PAR (1 << 3) /* ")" when "(" is open. */ #define ESC_SQU (1 << 4) /* "]" when "[" is open. */ #define ESC_FON (1 << 5) /* "*" immediately after unrelated "*". */ +#define ESC_EOL (1 << 6) /* " " at the and of a line. */ static int code_blocks, quote_blocks, list_blocks; static int outcount; @@ -307,8 +310,7 @@ md_node(struct roff_node *n) process_children = 1; n->flags &= ~NODE_ENDED; - switch (n->type) { - case ROFFT_TEXT: + if (n->type == ROFFT_TEXT) { if (n->flags & NODE_DELIMC) outflags &= ~(MD_spc | MD_spc_force); else if (outflags & MD_Sm) @@ -318,14 +320,21 @@ md_node(struct roff_node *n) outflags &= ~(MD_spc | MD_spc_force); else if (outflags & MD_Sm) outflags |= MD_spc; - break; - default: + } else if (n->tok < ROFF_MAX) { + switch (n->tok) { + case ROFF_br: + md_pre_br(n); + break; + default: + abort(); + } + } else { + assert(n->tok >= MDOC_Dd && n->tok < MDOC_MAX); act = md_acts + n->tok; cond = act->cond == NULL || (*act->cond)(n); if (cond && act->pre != NULL && (n->end == ENDBODY_NOT || n->child != NULL)) process_children = (*act->pre)(n); - break; } if (process_children && n->child != NULL) @@ -373,37 +382,43 @@ md_stack(char c) static void md_preword(void) { + const char *cp; + /* * If a list block is nested inside a code block or a blockquote, * blank lines for paragraph breaks no longer work; instead, * they terminate the list. Work around this markdown issue * by using mere line breaks instead. */ + if (list_blocks && outflags & MD_sp) { outflags &= ~MD_sp; outflags |= MD_br; } - /* End the old line if requested. */ + /* + * End the old line if requested. + * Escape whitespace at the end of the markdown line + * such that it won't look like an output line break. + */ if (outflags & MD_sp) putchar('\n'); else if (outflags & MD_br) { putchar(' '); putchar(' '); -#ifdef DEBUG - putchar(':'); - putchar(':'); - putchar(' '); - putchar(' '); -#endif - } + } else if (outflags & MD_nl && escflags & ESC_EOL) + md_named("zwnj"); /* Start a new line if necessary. */ if (outflags & (MD_nl | MD_br | MD_sp)) { putchar('\n'); - fputs(md_stack('\0'), stdout); + for (cp = md_stack('\0'); *cp != '\0'; cp++) { + putchar(*cp); + if (*cp == '>') + putchar(' '); + } outflags &= ~(MD_nl | MD_br | MD_sp); escflags = ESC_BOL; outcount = 0; @@ -436,7 +451,7 @@ md_rawword(const char *s) { md_preword(); - if (*s == 0) + if (*s == '\0') return; if (escflags & ESC_FON) { @@ -447,12 +462,6 @@ md_rawword(const char *s) while (*s != '\0') { switch(*s) { - case '(': - escflags |= ESC_PAR; - break; - case ')': - escflags |= ~ESC_PAR; - break; case '*': if (s[1] == '\0') escflags |= ESC_FON; @@ -469,6 +478,10 @@ md_rawword(const char *s) } md_char(*s++); } + if (s[-1] == ' ') + escflags |= ESC_EOL; + else + escflags &= ~ESC_EOL; } /* @@ -489,6 +502,9 @@ md_word(const char *s) md_preword(); + if (*s == '\0') + return; + /* No spacing after opening delimiters. */ if ((s[0] == '(' || s[0] == '[') && s[1] == '\0') outflags &= ~MD_spc; @@ -520,7 +536,7 @@ md_word(const char *s) bs = escflags & ESC_HYP && !code_blocks; break; case ')': - bs = escflags & ESC_PAR && !code_blocks; + bs = escflags & ESC_NUM && !code_blocks; break; case '*': case '[': @@ -629,7 +645,10 @@ md_word(const char *s) if (*currfont != '\0') { outflags &= ~MD_spc; md_rawword(currfont); - } + } else if (s[-2] == ' ') + escflags |= ESC_EOL; + else + escflags &= ~ESC_EOL; } /* @@ -639,7 +658,7 @@ static void md_named(const char *s) { printf("&%s;", s); - escflags &= ~ESC_FON; + escflags &= ~(ESC_FON | ESC_EOL); outcount++; } @@ -688,6 +707,8 @@ md_pre_raw(struct roff_node *n) if ((prefix = md_acts[n->tok].prefix) != NULL) { md_rawword(prefix); outflags &= ~MD_spc; + if (*prefix == '`') + code_blocks++; } return 1; } @@ -700,6 +721,8 @@ md_post_raw(struct roff_node *n) if ((suffix = md_acts[n->tok].suffix) != NULL) { outflags &= ~(MD_spc | MD_nl); md_rawword(suffix); + if (*suffix == '`') + code_blocks--; } } @@ -956,21 +979,17 @@ md_pre_Eo(struct roff_node *n) static void md_post_Eo(struct roff_node *n) { - int body, tail; - if (n->end != ENDBODY_NOT) { outflags |= MD_spc; return; } - body = n->child != NULL || n->parent->head->child != NULL; - tail = n->parent->tail != NULL && n->parent->tail->child != NULL; + if (n->child == NULL && n->parent->head->child == NULL) + return; - if (body && tail) + if (n->parent->tail != NULL && n->parent->tail->child != NULL) outflags &= ~MD_spc; - else if ( ! (body || tail)) - md_preword(); - else if ( ! tail) + else outflags |= MD_spc; } @@ -1018,6 +1037,15 @@ md_post_Fd(struct roff_node *n) outflags |= MD_br; } +static void +md_post_Fl(struct roff_node *n) +{ + md_post_raw(n); + if (n->child == NULL && n->next != NULL && + n->next->type != ROFFT_TEXT && !(n->next->flags & NODE_LINE)) + outflags &= ~MD_spc; +} + static int md_pre_Fn(struct roff_node *n) { @@ -1128,7 +1156,8 @@ md_pre_It(struct roff_node *n) case ROFFT_HEAD: bln = n->parent->parent; - if (bln->norm->Bl.comp == 0) + if (bln->norm->Bl.comp == 0 && + bln->norm->Bl.type != LIST_column) outflags |= MD_sp; outflags |= MD_nl; @@ -1154,9 +1183,14 @@ md_pre_It(struct roff_node *n) break; case LIST_enum: md_preword(); - printf("%d.\t", ++bln->norm->Bl.count); + if (bln->norm->Bl.count < 99) + bln->norm->Bl.count++; + printf("%d.\t", bln->norm->Bl.count); escflags &= ~ESC_FON; break; + case LIST_column: + outflags |= MD_br; + return 0; default: return 0; } @@ -1253,39 +1287,75 @@ md_post_Lb(struct roff_node *n) outflags |= MD_br; } +static void +md_uri(const char *s) +{ + while (*s != '\0') { + if (strchr("%()<>", *s) != NULL) { + printf("%%%2.2hhX", *s); + outcount += 3; + } else { + putchar(*s); + outcount++; + } + s++; + } +} + static int md_pre_Lk(struct roff_node *n) { const struct roff_node *link, *descr; - const unsigned char *s; if ((link = n->child) == NULL) return 0; - if ((descr = link->next) != NULL) { - md_rawword("["); - outflags &= ~MD_spc; - while (descr != NULL) { - md_word(descr->string); - descr = descr->next; - } - outflags &= ~MD_spc; - md_rawword("]("); - } else - md_rawword("<"); + /* Link text. */ + descr = link->next; + if (descr == NULL || descr->flags & NODE_DELIMC) + descr = link; /* no text */ + md_rawword("["); + outflags &= ~MD_spc; + do { + md_word(descr->string); + descr = descr->next; + } while (descr != NULL && !(descr->flags & NODE_DELIMC)); + outflags &= ~MD_spc; - for (s = link->string; *s != '\0'; s++) { - if (strchr("%)<>", *s) != NULL) { - printf("%%%2.2hhX", *s); - outcount += 3; - } else { - putchar(*s); + /* Link target. */ + md_rawword("]("); + md_uri(link->string); + outflags &= ~MD_spc; + md_rawword(")"); + + /* Trailing punctuation. */ + while (descr != NULL) { + md_word(descr->string); + descr = descr->next; + } + return 0; +} + +static int +md_pre_Mt(struct roff_node *n) +{ + const struct roff_node *nch; + + md_rawword("["); + outflags &= ~MD_spc; + for (nch = n->child; nch != NULL; nch = nch->next) + md_word(nch->string); + outflags &= ~MD_spc; + md_rawword("](mailto:"); + for (nch = n->child; nch != NULL; nch = nch->next) { + md_uri(nch->string); + if (nch->next != NULL) { + putchar(' '); outcount++; } } - outflags &= ~MD_spc; - md_rawword(link->next == NULL ? ">" : ")"); + md_rawword(")"); return 0; }