X-Git-Url: https://git.distorted.org.uk/~mdw/sgt/halibut/blobdiff_plain/7c8c4239f226d2dabdd84dabae9f5869d6d64f2f..206ec2a1f0384794b8ee3f3882acf40543c32cb0:/bk_pdf.c diff --git a/bk_pdf.c b/bk_pdf.c index f6babe5..6aba554 100644 --- a/bk_pdf.c +++ b/bk_pdf.c @@ -10,30 +10,7 @@ paragraph *pdf_config_filename(char *filename) { - paragraph *p; - wchar_t *ufilename, *up; - int len; - - p = mknew(paragraph); - memset(p, 0, sizeof(*p)); - p->type = para_Config; - p->next = NULL; - p->fpos.filename = ""; - p->fpos.line = p->fpos.col = -1; - - ufilename = ufroma_dup(filename); - len = ustrlen(ufilename) + 2 + lenof(L"pdf-filename"); - p->keyword = mknewa(wchar_t, len); - up = p->keyword; - ustrcpy(up, L"pdf-filename"); - up = uadv(up); - ustrcpy(up, ufilename); - up = uadv(up); - *up = L'\0'; - assert(up - p->keyword < len); - sfree(ufilename); - - return p; + return cmdline_cfg_simple("pdf-filename", filename, NULL); } typedef struct object_Tag object; @@ -58,6 +35,8 @@ static void objtext(object *o, char const *text); static void objstream(object *o, char const *text); static void pdf_string(void (*add)(object *, char const *), object *, char const *); +static void pdf_string_len(void (*add)(object *, char const *), + object *, char const *, int); static void objref(object *o, object *dest); static void make_pages_node(object *node, object *parent, page_data *first, @@ -88,7 +67,7 @@ void pdf_backend(paragraph *sourceform, keywordlist *keywords, if (p->type == para_Config && p->parent) { if (!ustricmp(p->keyword, L"pdf-filename")) { sfree(filename); - filename = utoa_dup(uadv(p->keyword)); + filename = dupstr(adv(p->origkeyword)); } } } @@ -97,7 +76,10 @@ void pdf_backend(paragraph *sourceform, keywordlist *keywords, olist.number = 1; cat = new_object(&olist); - outlines = new_object(&olist); + if (doc->n_outline_elements > 0) + outlines = new_object(&olist); + else + outlines = NULL; pages = new_object(&olist); resources = new_object(&olist); @@ -105,11 +87,16 @@ void pdf_backend(paragraph *sourceform, keywordlist *keywords, * The catalogue just contains references to the outlines and * pages objects. */ - objtext(cat, "<<\n/Type /Catalog\n/Outlines "); - objref(cat, outlines); + objtext(cat, "<<\n/Type /Catalog"); + if (outlines) { + objtext(cat, "\n/Outlines "); + objref(cat, outlines); + } objtext(cat, "\n/Pages "); objref(cat, pages); - objtext(cat, "\n/PageMode /UseOutlines\n>>\n"); + if (outlines) + objtext(cat, "\n/PageMode /UseOutlines"); + objtext(cat, "\n>>\n"); /* * Set up the resources dictionary, which mostly means @@ -370,7 +357,7 @@ void pdf_backend(paragraph *sourceform, keywordlist *keywords, /* * Set up the outlines dictionary. */ - { + if (outlines) { int topcount; char buf[80]; @@ -477,7 +464,7 @@ void pdf_backend(paragraph *sourceform, keywordlist *keywords, static object *new_object(objlist *list) { - object *obj = mknew(object); + object *obj = snew(object); obj->list = list; @@ -603,44 +590,32 @@ static void make_pages_node(object *node, object *parent, page_data *first, * means that glyphs are identified by PS strings and hence font * encoding can be managed independently of the supplied encoding * of the font. However, in the document outline, the PDF spec - * simply asks for ordinary text strings without mentioning what - * character set they are supposed to be interpreted in. - * - * Therefore, for the moment, I'm going to assume they're US-ASCII - * only. If anyone knows better, they should let me know :-/ + * encodes in either PDFDocEncoding (a custom superset of + * ISO-8859-1) or UTF-16BE. */ -static int pdf_convert(wchar_t *s, char **result) { - int doing = (result != 0); - int ok = TRUE; - char *p = NULL; - int plen = 0, psize = 0; - - for (; *s; s++) { - wchar_t c = *s; - char outc; - - if (c >= 32 && c <= 126) { - /* Char is OK. */ - outc = (char)c; - } else { - /* Char is not OK. */ - ok = FALSE; - outc = 0xBF; /* approximate the good old DEC `uh?' */ - } - if (doing) { - if (plen >= psize) { - psize = plen + 256; - p = resize(p, psize); - } - p[plen++] = outc; - } +static char *pdf_outline_convert(wchar_t *s, int *len) { + char *ret; + + ret = utoa_careful_dup(s, CS_PDF); + + /* + * Very silly special case: if the returned string begins with + * FE FF, then the PDF reader will mistake it for a UTF-16BE + * string. So in this case we give up on PDFDocEncoding and + * encode it in UTF-16 straight away. + */ + if (ret && ret[0] == '\xFE' && ret[1] == '\xFF') { + sfree(ret); + ret = NULL; } - if (doing) { - p = resize(p, plen+1); - p[plen] = '\0'; - *result = p; + + if (!ret) { + ret = utoa_dup_len(s, CS_UTF16BE, len); + } else { + *len = strlen(ret); } - return ok; + + return ret; } static int make_outline(object *parent, outline_element *items, int n, @@ -656,6 +631,7 @@ static int make_outline(object *parent, outline_element *items, int n, while (n > 0) { char *title; + int titlelen; /* * Here we expect to be sitting on an item at the given @@ -664,14 +640,14 @@ static int make_outline(object *parent, outline_element *items, int n, */ assert(items->level == level); - pdf_convert(items->pdata->outline_title, &title); + title = pdf_outline_convert(items->pdata->outline_title, &titlelen); totalcount++; curr = new_object(parent->list); if (!first) first = curr; last = curr; objtext(curr, "<<\n/Title "); - pdf_string(objtext, curr, title); + pdf_string_len(objtext, curr, title, titlelen); objtext(curr, "\n/Parent "); objref(curr, parent); objtext(curr, "\n/Dest ["); @@ -742,7 +718,7 @@ static int pdf_versionid(FILE *fp, word *words) switch (type) { case word_Normal: - text = utoa_dup(words->text); + text = utoa_dup(words->text, CS_ASCII); break; case word_WhiteSpace: text = dupstr(" "); @@ -762,19 +738,30 @@ static int pdf_versionid(FILE *fp, word *words) return ret; } -static void pdf_string(void (*add)(object *, char const *), - object *o, char const *str) +static void pdf_string_len(void (*add)(object *, char const *), + object *o, char const *str, int len) { char const *p; add(o, "("); - for (p = str; *p; p++) { - char c[2]; - if (*p == '\\' || *p == '(' || *p == ')') - add(o, "\\"); - c[0] = *p; - c[1] = '\0'; + for (p = str; len > 0; p++, len--) { + char c[10]; + if (*p < ' ' || *p > '~') { + sprintf(c, "\\%03o", 0xFF & (int)*p); + } else { + int n = 0; + if (*p == '\\' || *p == '(' || *p == ')') + c[n++] = '\\'; + c[n++] = *p; + c[n] = '\0'; + } add(o, c); } add(o, ")"); } + +static void pdf_string(void (*add)(object *, char const *), + object *o, char const *str) +{ + pdf_string_len(add, o, str, strlen(str)); +}