=================================================================== RCS file: /cvs/mandoc/mdoc_html.c,v retrieving revision 1.322 retrieving revision 1.325 diff -u -p -r1.322 -r1.325 --- mandoc/mdoc_html.c 2018/12/31 10:35:56 1.322 +++ mandoc/mdoc_html.c 2019/01/11 12:56:43 1.325 @@ -1,4 +1,4 @@ -/* $Id: mdoc_html.c,v 1.322 2018/12/31 10:35:56 schwarze Exp $ */ +/* $Id: mdoc_html.c,v 1.325 2019/01/11 12:56:43 schwarze Exp $ */ /* * Copyright (c) 2008-2011, 2014 Kristaps Dzonsons * Copyright (c) 2014,2015,2016,2017,2018 Ingo Schwarze @@ -268,18 +268,17 @@ synopsis_pre(struct html *h, const struct roff_node *n case MDOC_Fo: case MDOC_In: case MDOC_Vt: - print_paragraph(h); break; case MDOC_Ft: - if (MDOC_Fn != n->tok && MDOC_Fo != n->tok) { - print_paragraph(h); + if (n->tok != MDOC_Fn && n->tok != MDOC_Fo) break; - } /* FALLTHROUGH */ default: print_otag(h, TAG_BR, ""); - break; + return; } + html_close_paragraph(h); + print_otag(h, TAG_P, "c", "Pp"); } void @@ -346,16 +345,20 @@ print_mdoc_nodelist(MDOC_ARGS) static void print_mdoc_node(MDOC_ARGS) { - int child; struct tag *t; + int child; if (n->type == ROFFT_COMMENT || n->flags & NODE_NOPRT) return; + html_fillmode(h, n->flags & NODE_NOFILL ? ROFF_nf : ROFF_fi); + child = 1; t = h->tag; - n->flags &= ~NODE_ENDED; + if (t->tag == TAG_P || t->tag == TAG_PRE) + t = t->next; + n->flags &= ~NODE_ENDED; switch (n->type) { case ROFFT_TEXT: /* No tables in this mode... */ @@ -374,7 +377,7 @@ print_mdoc_node(MDOC_ARGS) print_text(h, n->string); if (NODE_DELIMO & n->flags) h->flags |= HTML_NOSPACE; - return; + break; case ROFFT_EQN: print_eqn(h, n->eqn); break; @@ -399,8 +402,8 @@ print_mdoc_node(MDOC_ARGS) assert(h->tblt == NULL); if (n->tok < ROFF_MAX) { roff_html_pre(h, n); - child = 0; - break; + print_stagq(h, t); + return; } assert(n->tok >= MDOC_Dd && n->tok < MDOC_MAX); if (mdoc_html_acts[n->tok - MDOC_Dd].pre != NULL && @@ -415,17 +418,17 @@ print_mdoc_node(MDOC_ARGS) h->flags |= HTML_PREKEEP; } - if (child && n->child) + if (child && n->child != NULL) print_mdoc_nodelist(meta, n->child, h); print_stagq(h, t); switch (n->type) { + case ROFFT_TEXT: case ROFFT_EQN: break; default: - if (n->tok < ROFF_MAX || - mdoc_html_acts[n->tok - MDOC_Dd].post == NULL || + if (mdoc_html_acts[n->tok - MDOC_Dd].post == NULL || n->flags & NODE_ENDED) break; (*mdoc_html_acts[n->tok - MDOC_Dd].post)(meta, n, h); @@ -433,6 +436,12 @@ print_mdoc_node(MDOC_ARGS) n->body->flags |= NODE_ENDED; break; } + + if (n->flags & NODE_NOFILL && + (n->next == NULL || n->next->flags & NODE_LINE)) { + h->col++; + print_endline(h); + } } static void @@ -516,6 +525,7 @@ mdoc_sh_pre(MDOC_ARGS) switch (n->type) { case ROFFT_BLOCK: + html_close_paragraph(h); if ((h->oflags & HTML_TOC) == 0 || h->flags & HTML_TOCDONE || n->sec <= SEC_SYNOPSIS) @@ -563,7 +573,7 @@ mdoc_sh_pre(MDOC_ARGS) break; case ROFFT_HEAD: id = html_make_id(n, 1); - print_otag(h, TAG_H1, "cTi", "Sh", id); + print_otag(h, TAG_H1, "ci", "Sh", id); if (id != NULL) print_otag(h, TAG_A, "chR", "permalink", id); break; @@ -582,11 +592,20 @@ mdoc_ss_pre(MDOC_ARGS) { char *id; - if (n->type != ROFFT_HEAD) + switch (n->type) { + case ROFFT_BLOCK: + html_close_paragraph(h); return 1; + case ROFFT_HEAD: + break; + case ROFFT_BODY: + return 1; + default: + abort(); + } id = html_make_id(n, 1); - print_otag(h, TAG_H2, "cTi", "Ss", id); + print_otag(h, TAG_H2, "ci", "Ss", id); if (id != NULL) print_otag(h, TAG_A, "chR", "permalink", id); return 1; @@ -599,7 +618,7 @@ mdoc_fl_pre(MDOC_ARGS) if ((id = cond_id(n)) != NULL) print_otag(h, TAG_A, "chR", "permalink", id); - print_otag(h, TAG_CODE, "cTi", "Fl", id); + print_otag(h, TAG_CODE, "ci", "Fl", id); print_text(h, "\\-"); if (!(n->child == NULL && @@ -618,19 +637,27 @@ mdoc_cm_pre(MDOC_ARGS) if ((id = cond_id(n)) != NULL) print_otag(h, TAG_A, "chR", "permalink", id); - print_otag(h, TAG_CODE, "cTi", "Cm", id); + print_otag(h, TAG_CODE, "ci", "Cm", id); return 1; } static int mdoc_nd_pre(MDOC_ARGS) { - if (n->type != ROFFT_BODY) + switch (n->type) { + case ROFFT_BLOCK: + html_close_paragraph(h); return 1; - + case ROFFT_HEAD: + return 0; + case ROFFT_BODY: + break; + default: + abort(); + } print_text(h, "\\(em"); /* Cannot use TAG_SPAN because it may contain blocks. */ - print_otag(h, TAG_DIV, "cT", "Nd"); + print_otag(h, TAG_DIV, "c", "Nd"); return 1; } @@ -638,18 +665,21 @@ static int mdoc_nm_pre(MDOC_ARGS) { switch (n->type) { + case ROFFT_BLOCK: + break; case ROFFT_HEAD: print_otag(h, TAG_TD, ""); /* FALLTHROUGH */ case ROFFT_ELEM: - print_otag(h, TAG_CODE, "cT", "Nm"); + print_otag(h, TAG_CODE, "c", "Nm"); return 1; case ROFFT_BODY: print_otag(h, TAG_TD, ""); return 1; default: - break; + abort(); } + html_close_paragraph(h); synopsis_pre(h, n); print_otag(h, TAG_TABLE, "c", "Nm"); print_otag(h, TAG_TR, ""); @@ -663,11 +693,11 @@ mdoc_xr_pre(MDOC_ARGS) return 0; if (h->base_man1) - print_otag(h, TAG_A, "cThM", "Xr", + print_otag(h, TAG_A, "chM", "Xr", n->child->string, n->child->next == NULL ? NULL : n->child->next->string); else - print_otag(h, TAG_A, "cT", "Xr"); + print_otag(h, TAG_A, "c", "Xr"); n = n->child; print_text(h, n->string); @@ -696,7 +726,7 @@ mdoc_ns_pre(MDOC_ARGS) static int mdoc_ar_pre(MDOC_ARGS) { - print_otag(h, TAG_VAR, "cT", "Ar"); + print_otag(h, TAG_VAR, "c", "Ar"); return 1; } @@ -711,7 +741,6 @@ static int mdoc_it_pre(MDOC_ARGS) { const struct roff_node *bl; - struct tag *t; enum mdoc_list type; bl = n->parent; @@ -753,17 +782,6 @@ mdoc_it_pre(MDOC_ARGS) case LIST_tag: switch (n->type) { case ROFFT_HEAD: - if (h->style != NULL && !bl->norm->Bl.comp && - (n->parent->prev == NULL || - n->parent->prev->body == NULL || - n->parent->prev->body->child != NULL)) { - t = print_otag(h, TAG_DT, ""); - print_text(h, "\\ "); - print_tagq(h, t); - t = print_otag(h, TAG_DD, ""); - print_text(h, "\\ "); - print_tagq(h, t); - } print_otag(h, TAG_DT, ""); break; case ROFFT_BODY: @@ -802,12 +820,15 @@ mdoc_bl_pre(MDOC_ARGS) enum htmltag elemtype; switch (n->type) { - case ROFFT_BODY: - return 1; + case ROFFT_BLOCK: + html_close_paragraph(h); + break; case ROFFT_HEAD: return 0; + case ROFFT_BODY: + return 1; default: - break; + abort(); } bl = &n->norm->Bl; @@ -877,28 +898,34 @@ mdoc_ex_pre(MDOC_ARGS) static int mdoc_st_pre(MDOC_ARGS) { - print_otag(h, TAG_SPAN, "cT", "St"); + print_otag(h, TAG_SPAN, "c", "St"); return 1; } static int mdoc_em_pre(MDOC_ARGS) { - print_otag(h, TAG_I, "cT", "Em"); + print_otag(h, TAG_I, "c", "Em"); return 1; } static int mdoc_d1_pre(MDOC_ARGS) { - if (n->type != ROFFT_BLOCK) + switch (n->type) { + case ROFFT_BLOCK: + html_close_paragraph(h); + break; + case ROFFT_HEAD: + return 0; + case ROFFT_BODY: return 1; - + default: + abort(); + } print_otag(h, TAG_DIV, "c", "Bd Bd-indent"); - if (n->tok == MDOC_Dl) print_otag(h, TAG_CODE, "c", "Li"); - return 1; } @@ -908,7 +935,7 @@ mdoc_sx_pre(MDOC_ARGS) char *id; id = html_make_id(n, 0); - print_otag(h, TAG_A, "cThR", "Sx", id); + print_otag(h, TAG_A, "chR", "Sx", id); free(id); return 1; } @@ -916,75 +943,51 @@ mdoc_sx_pre(MDOC_ARGS) static int mdoc_bd_pre(MDOC_ARGS) { - int comp; + char buf[16]; struct roff_node *nn; + int comp; - if (n->type == ROFFT_HEAD) - return 0; - - if (n->type == ROFFT_BLOCK) { - comp = n->norm->Bd.comp; - for (nn = n; nn && ! comp; nn = nn->parent) { - if (nn->type != ROFFT_BLOCK) - continue; - if (MDOC_Ss == nn->tok || MDOC_Sh == nn->tok) - comp = 1; - if (nn->prev) - break; - } - if ( ! comp) - print_paragraph(h); + switch (n->type) { + case ROFFT_BLOCK: + html_close_paragraph(h); return 1; + case ROFFT_HEAD: + return 0; + case ROFFT_BODY: + break; + default: + abort(); } - /* Handle the -offset argument. */ + /* Handle preceding whitespace. */ - if (n->norm->Bd.offs == NULL || - ! strcmp(n->norm->Bd.offs, "left")) - print_otag(h, TAG_DIV, "c", "Bd"); - else - print_otag(h, TAG_DIV, "c", "Bd Bd-indent"); - - if (n->norm->Bd.type != DISP_unfilled && - n->norm->Bd.type != DISP_literal) - return 1; - - print_otag(h, TAG_PRE, "c", "Li"); - for (nn = n->child; nn; nn = nn->next) { - print_mdoc_node(meta, nn, h); - /* - * If the printed node flushes its own line, then we - * needn't do it here as well. This is hacky, but the - * notion of selective eoln whitespace is pretty dumb - * anyway, so don't sweat it. - */ - switch (nn->tok) { - case ROFF_br: - case ROFF_sp: - case MDOC_Sm: - case MDOC_Bl: - case MDOC_D1: - case MDOC_Dl: - case MDOC_Pp: + comp = n->norm->Bd.comp; + for (nn = n; nn != NULL && comp == 0; nn = nn->parent) { + if (nn->type != ROFFT_BLOCK) continue; - default: + if (nn->tok == MDOC_Sh || nn->tok == MDOC_Ss) + comp = 1; + if (nn->prev != NULL) break; - } - if (h->flags & HTML_NONEWLINE || - (nn->next && ! (nn->next->flags & NODE_LINE))) - continue; - else if (nn->next) - print_text(h, "\n"); - - h->flags |= HTML_NOSPACE; } - return 0; + (void)strlcpy(buf, "Bd", sizeof(buf)); + if (comp == 0) + (void)strlcat(buf, " Pp", sizeof(buf)); + + /* Handle the -offset argument. */ + + if (n->norm->Bd.offs != NULL && + strcmp(n->norm->Bd.offs, "left") != 0) + (void)strlcat(buf, " Bd-indent", sizeof(buf)); + + print_otag(h, TAG_DIV, "c", buf); + return 1; } static int mdoc_pa_pre(MDOC_ARGS) { - print_otag(h, TAG_SPAN, "cT", "Pa"); + print_otag(h, TAG_SPAN, "c", "Pa"); return 1; } @@ -1015,7 +1018,7 @@ mdoc_an_pre(MDOC_ARGS) if (n->sec == SEC_AUTHORS && ! (h->flags & HTML_NOSPLIT)) h->flags |= HTML_SPLIT; - print_otag(h, TAG_SPAN, "cT", "An"); + print_otag(h, TAG_SPAN, "c", "An"); return 1; } @@ -1023,7 +1026,7 @@ static int mdoc_cd_pre(MDOC_ARGS) { synopsis_pre(h, n); - print_otag(h, TAG_CODE, "cT", "Cd"); + print_otag(h, TAG_CODE, "c", "Cd"); return 1; } @@ -1034,7 +1037,7 @@ mdoc_dv_pre(MDOC_ARGS) if ((id = cond_id(n)) != NULL) print_otag(h, TAG_A, "chR", "permalink", id); - print_otag(h, TAG_CODE, "cTi", "Dv", id); + print_otag(h, TAG_CODE, "ci", "Dv", id); return 1; } @@ -1045,7 +1048,7 @@ mdoc_ev_pre(MDOC_ARGS) if ((id = cond_id(n)) != NULL) print_otag(h, TAG_A, "chR", "permalink", id); - print_otag(h, TAG_CODE, "cTi", "Ev", id); + print_otag(h, TAG_CODE, "ci", "Ev", id); return 1; } @@ -1062,7 +1065,7 @@ mdoc_er_pre(MDOC_ARGS) if (id != NULL) print_otag(h, TAG_A, "chR", "permalink", id); - print_otag(h, TAG_CODE, "cTi", "Er", id); + print_otag(h, TAG_CODE, "ci", "Er", id); return 1; } @@ -1073,12 +1076,12 @@ mdoc_fa_pre(MDOC_ARGS) struct tag *t; if (n->parent->tok != MDOC_Fo) { - print_otag(h, TAG_VAR, "cT", "Fa"); + print_otag(h, TAG_VAR, "c", "Fa"); return 1; } for (nn = n->child; nn; nn = nn->next) { - t = print_otag(h, TAG_VAR, "cT", "Fa"); + t = print_otag(h, TAG_VAR, "c", "Fa"); print_text(h, nn->string); print_tagq(h, t); if (nn->next) { @@ -1109,11 +1112,11 @@ mdoc_fd_pre(MDOC_ARGS) assert(n->type == ROFFT_TEXT); if (strcmp(n->string, "#include")) { - print_otag(h, TAG_CODE, "cT", "Fd"); + print_otag(h, TAG_CODE, "c", "Fd"); return 1; } - print_otag(h, TAG_CODE, "cT", "In"); + print_otag(h, TAG_CODE, "c", "In"); print_text(h, n->string); if (NULL != (n = n->next)) { @@ -1127,10 +1130,10 @@ mdoc_fd_pre(MDOC_ARGS) cp = strchr(buf, '\0') - 1; if (cp >= buf && (*cp == '>' || *cp == '"')) *cp = '\0'; - t = print_otag(h, TAG_A, "cThI", "In", buf); + t = print_otag(h, TAG_A, "chI", "In", buf); free(buf); } else - t = print_otag(h, TAG_A, "cT", "In"); + t = print_otag(h, TAG_A, "c", "In"); print_text(h, n->string); print_tagq(h, t); @@ -1157,7 +1160,7 @@ mdoc_vt_pre(MDOC_ARGS) } else if (n->type == ROFFT_HEAD) return 0; - print_otag(h, TAG_VAR, "cT", "Vt"); + print_otag(h, TAG_VAR, "c", "Vt"); return 1; } @@ -1165,7 +1168,7 @@ static int mdoc_ft_pre(MDOC_ARGS) { synopsis_pre(h, n); - print_otag(h, TAG_VAR, "cT", "Ft"); + print_otag(h, TAG_VAR, "c", "Ft"); return 1; } @@ -1186,7 +1189,7 @@ mdoc_fn_pre(MDOC_ARGS) ep = strchr(sp, ' '); if (NULL != ep) { - t = print_otag(h, TAG_VAR, "cT", "Ft"); + t = print_otag(h, TAG_VAR, "c", "Ft"); while (ep) { sz = MIN((int)(ep - sp), BUFSIZ - 1); @@ -1199,7 +1202,7 @@ mdoc_fn_pre(MDOC_ARGS) print_tagq(h, t); } - t = print_otag(h, TAG_CODE, "cT", "Fn"); + t = print_otag(h, TAG_CODE, "c", "Fn"); if (sp) print_text(h, sp); @@ -1212,10 +1215,10 @@ mdoc_fn_pre(MDOC_ARGS) for (n = n->child->next; n; n = n->next) { if (NODE_SYNPRETTY & n->flags) - t = print_otag(h, TAG_VAR, "cTs", "Fa", + t = print_otag(h, TAG_VAR, "cs", "Fa", "white-space", "nowrap"); else - t = print_otag(h, TAG_VAR, "cT", "Fa"); + t = print_otag(h, TAG_VAR, "c", "Fa"); print_text(h, n->string); print_tagq(h, t); if (n->next) { @@ -1262,8 +1265,10 @@ mdoc_skip_pre(MDOC_ARGS) static int mdoc_pp_pre(MDOC_ARGS) { - - print_paragraph(h); + if ((n->flags & NODE_NOFILL) == 0) { + html_close_paragraph(h); + print_otag(h, TAG_P, "c", "Pp"); + } return 0; } @@ -1286,7 +1291,7 @@ mdoc_lk_pre(MDOC_ARGS) descr = link->next; if (descr == punct) descr = link; /* no text */ - t = print_otag(h, TAG_A, "cTh", "Lk", link->string); + t = print_otag(h, TAG_A, "ch", "Lk", link->string); do { if (descr->flags & (NODE_DELIMC | NODE_DELIMO)) h->flags |= HTML_NOSPACE; @@ -1314,7 +1319,7 @@ mdoc_mt_pre(MDOC_ARGS) assert(n->type == ROFFT_TEXT); mandoc_asprintf(&cp, "mailto:%s", n->string); - t = print_otag(h, TAG_A, "cTh", "Mt", cp); + t = print_otag(h, TAG_A, "ch", "Mt", cp); print_text(h, n->string); print_tagq(h, t); free(cp); @@ -1342,7 +1347,7 @@ mdoc_fo_pre(MDOC_ARGS) return 0; assert(n->child->string); - t = print_otag(h, TAG_CODE, "cT", "Fn"); + t = print_otag(h, TAG_CODE, "c", "Fn"); print_text(h, n->child->string); print_tagq(h, t); return 0; @@ -1366,7 +1371,7 @@ mdoc_in_pre(MDOC_ARGS) struct tag *t; synopsis_pre(h, n); - print_otag(h, TAG_CODE, "cT", "In"); + print_otag(h, TAG_CODE, "c", "In"); /* * The first argument of the `In' gets special treatment as @@ -1385,9 +1390,9 @@ mdoc_in_pre(MDOC_ARGS) assert(n->type == ROFFT_TEXT); if (h->base_includes) - t = print_otag(h, TAG_A, "cThI", "In", n->string); + t = print_otag(h, TAG_A, "chI", "In", n->string); else - t = print_otag(h, TAG_A, "cT", "In"); + t = print_otag(h, TAG_A, "c", "In"); print_text(h, n->string); print_tagq(h, t); @@ -1412,14 +1417,14 @@ mdoc_ic_pre(MDOC_ARGS) if ((id = cond_id(n)) != NULL) print_otag(h, TAG_A, "chR", "permalink", id); - print_otag(h, TAG_CODE, "cTi", "Ic", id); + print_otag(h, TAG_CODE, "ci", "Ic", id); return 1; } static int mdoc_va_pre(MDOC_ARGS) { - print_otag(h, TAG_VAR, "cT", "Va"); + print_otag(h, TAG_VAR, "c", "Va"); return 1; } @@ -1438,10 +1443,17 @@ mdoc_bf_pre(MDOC_ARGS) { const char *cattr; - if (n->type == ROFFT_HEAD) - return 0; - else if (n->type != ROFFT_BODY) + switch (n->type) { + case ROFFT_BLOCK: + html_close_paragraph(h); return 1; + case ROFFT_HEAD: + return 0; + case ROFFT_BODY: + break; + default: + abort(); + } if (FONT_Em == n->norm->Bf.font) cattr = "Bf Em"; @@ -1464,7 +1476,7 @@ mdoc_ms_pre(MDOC_ARGS) if ((id = cond_id(n)) != NULL) print_otag(h, TAG_A, "chR", "permalink", id); - print_otag(h, TAG_SPAN, "cTi", "Ms", id); + print_otag(h, TAG_SPAN, "ci", "Ms", id); return 1; } @@ -1487,13 +1499,21 @@ mdoc_pf_post(MDOC_ARGS) static int mdoc_rs_pre(MDOC_ARGS) { - if (n->type != ROFFT_BLOCK) - return 1; - - if (n->prev && SEC_SEE_ALSO == n->sec) - print_paragraph(h); - - print_otag(h, TAG_CITE, "cT", "Rs"); + switch (n->type) { + case ROFFT_BLOCK: + if (n->sec == SEC_SEE_ALSO) + html_close_paragraph(h); + break; + case ROFFT_HEAD: + return 0; + case ROFFT_BODY: + if (n->sec == SEC_SEE_ALSO) + print_otag(h, TAG_P, "c", "Pp"); + print_otag(h, TAG_CITE, "c", "Rs"); + break; + default: + abort(); + } return 1; } @@ -1522,7 +1542,7 @@ mdoc_li_pre(MDOC_ARGS) static int mdoc_sy_pre(MDOC_ARGS) { - print_otag(h, TAG_B, "cT", "Sy"); + print_otag(h, TAG_B, "c", "Sy"); return 1; } @@ -1532,7 +1552,7 @@ mdoc_lb_pre(MDOC_ARGS) if (SEC_LIBRARY == n->sec && NODE_LINE & n->flags && n->prev) print_otag(h, TAG_BR, ""); - print_otag(h, TAG_SPAN, "cT", "Lb"); + print_otag(h, TAG_SPAN, "c", "Lb"); return 1; } @@ -1670,9 +1690,15 @@ mdoc_quote_pre(MDOC_ARGS) case MDOC_Oo: case MDOC_Op: print_text(h, "\\(lB"); - h->flags |= HTML_NOSPACE; - /* Cannot use TAG_SPAN because it may contain blocks. */ - print_otag(h, TAG_IDIV, "c", "Op"); + /* + * Give up on semantic markup for now. + * We cannot use TAG_SPAN because .Oo may contain blocks. + * We cannot use TAG_IDIV because we might be in a + * phrasing context (like .Dl or .Pp); we cannot + * close out a .Pp at this point either because + * that would break the line. + */ + /* XXX print_otag(h, TAG_???, "c", "Op"); */ break; case MDOC_En: if (NULL == n->norm->Es ||