/*
- * Copyright © 2008 Red Hat, Inc.
+ * Copyright © 2008,2009 Red Hat, Inc.
*
* Red Hat Author(s): Behdad Esfahbod
*
#include <stdarg.h>
+/* XXX Document the language.
+ *
+ * These are mostly the features implemented but not documented:
+ *
+ * width %[[-]width]{tag}
+ * index %{tag[ids]}
+ * name= %{tag=|decorator}
+ * :name= %{:tag=|decorator}
+ * subexpr %{{expr}|decorator1|decorator2}
+ * delete %{-charset,lang{expr}|decorator}
+ * filter %{+family,familylang{expr}|decorator}
+ * cond %{?tag1,tag2,!tag3{}{}}
+ * decorat %{tag|decorator1|decorator2}
+ * default %{parameter:-word}
+ * array %{[]family,familylang{expr}|decorator}
+ * langset enumeration using the same syntax as arrays
+ *
+ * filters:
+ * basename FcStrBasename
+ * dirname FcStrDirname
+ * downcase FcStrDowncase
+ * shescape
+ * cescape
+ * xmlescape
+ * delete delete chars
+ * escape escape chars
+ * translate translate chars
+ *
+ * builtins:
+ * unparse
+ * fcmatch
+ * fclist
+ * pkgkit
+ */
+
/*
* Some ideas for future syntax extensions:
*
* - allow indexing subexprs using '%{[idx]elt1,elt2{subexpr}}'
- * - allow indexing simple tags using '%{elt[idx]}'
- * - conditional/filtering/deletion on binding (using '(w)'/'(s)' notation)
+ * - conditional/filtering/deletion on binding (using '(w)'/'(s)'/'(=)' notation)
*/
+
+#define FCMATCH_FORMAT "%{file:-<unknown filename>|basename}: \"%{family[0]:-<unknown family>}\" \"%{style[0]:-<unknown style>}\""
+#define FCLIST_FORMAT "%{?file{%{file}: }}%{=unparse}"
+#define PKGKIT_FORMAT "%{[]family{font(%{family|downcase|delete( )})\n}}%{[]lang{font(:lang=%{lang|downcase|translate(_,-)})\n}}"
+
+
static void
message (const char *fmt, ...)
{
const FcChar8 *format;
int format_len;
FcChar8 *word;
+ FcBool word_allocated;
} FcFormatContext;
static FcBool
FcFormatContextInit (FcFormatContext *c,
- const FcChar8 *format)
+ const FcChar8 *format,
+ FcChar8 *scratch,
+ int scratch_len)
{
c->format_orig = c->format = format;
c->format_len = strlen ((const char *) format);
- c->word = malloc (c->format_len + 1);
+
+ if (c->format_len < scratch_len)
+ {
+ c->word = scratch;
+ c->word_allocated = FcFalse;
+ }
+ else
+ {
+ c->word = malloc (c->format_len + 1);
+ c->word_allocated = FcTrue;
+ }
return c->word != NULL;
}
static void
FcFormatContextDone (FcFormatContext *c)
{
- if (c)
+ if (c && c->word_allocated)
{
free (c->word);
}
return FcFalse;
}
+static char escaped_char(const char ch)
+{
+ switch (ch) {
+ case 'a': return '\a';
+ case 'b': return '\b';
+ case 'f': return '\f';
+ case 'n': return '\n';
+ case 'r': return '\r';
+ case 't': return '\t';
+ case 'v': return '\v';
+ default: return ch;
+ }
+}
+
static FcBool
read_word (FcFormatContext *c)
{
{
c->format++;
if (*c->format)
- c->format++;
+ *p++ = escaped_char (*c->format++);
continue;
}
else if (FcCharIsPunct (*c->format))
if (p == c->word)
{
- message ("expected element name at %d",
+ message ("expected identifier at %d",
c->format - c->format_orig + 1);
return FcFalse;
}
return FcTrue;
}
+static FcBool
+read_chars (FcFormatContext *c,
+ FcChar8 term)
+{
+ FcChar8 *p;
+
+ p = c->word;
+
+ while (*c->format && *c->format != '}' && *c->format != term)
+ {
+ if (*c->format == '\\')
+ {
+ c->format++;
+ if (*c->format)
+ *p++ = escaped_char (*c->format++);
+ continue;
+ }
+
+ *p++ = *c->format++;
+ }
+ *p = '\0';
+
+ if (p == c->word)
+ {
+ message ("expected character data at %d",
+ c->format - c->format_orig + 1);
+ return FcFalse;
+ }
+
+ return FcTrue;
+}
+
+static FcBool
+FcPatternFormatToBuf (FcPattern *pat,
+ const FcChar8 *format,
+ FcStrBuf *buf);
+
+static FcBool
+interpret_builtin (FcFormatContext *c,
+ FcPattern *pat,
+ FcStrBuf *buf)
+{
+ FcChar8 *new_str;
+ FcBool ret;
+
+ if (!expect_char (c, '=') ||
+ !read_word (c))
+ return FcFalse;
+
+ /* try simple builtins first */
+ if (0) { }
+#define BUILTIN(name, func) \
+ else if (0 == strcmp ((const char *) c->word, name))\
+ do { new_str = func (pat); ret = FcTrue; } while (0)
+ BUILTIN ("unparse", FcNameUnparse);
+ /* BUILTIN ("verbose", FcPatternPrint); XXX */
+#undef BUILTIN
+ else
+ ret = FcFalse;
+
+ if (ret)
+ {
+ if (new_str)
+ {
+ FcStrBufString (buf, new_str);
+ free (new_str);
+ return FcTrue;
+ }
+ else
+ return FcFalse;
+ }
+
+ /* now try our custom formats */
+ if (0) { }
+#define BUILTIN(name, format) \
+ else if (0 == strcmp ((const char *) c->word, name))\
+ ret = FcPatternFormatToBuf (pat, (const FcChar8 *) format, buf)
+ BUILTIN ("fcmatch", FCMATCH_FORMAT);
+ BUILTIN ("fclist", FCLIST_FORMAT);
+ BUILTIN ("pkgkit", PKGKIT_FORMAT);
+#undef BUILTIN
+ else
+ ret = FcFalse;
+
+ if (!ret)
+ message ("unknown builtin \"%s\"",
+ c->word);
+
+ return ret;
+}
+
static FcBool
interpret_expr (FcFormatContext *c,
FcPattern *pat,
FcTrue;
}
-static FcBool
-interpret_simple (FcFormatContext *c,
- FcPattern *pat,
- FcStrBuf *buf)
-{
- FcPatternElt *e;
- FcBool add_colon = FcFalse;
- FcBool add_elt_name = FcFalse;
-
- if (consume_char (c, ':'))
- add_colon = FcTrue;
-
- if (!read_word (c))
- return FcFalse;
-
- if (consume_char (c, '='))
- add_elt_name = FcTrue;
-
- e = FcPatternObjectFindElt (pat,
- FcObjectFromName ((const char *) c->word));
- if (e)
- {
- FcValueListPtr l;
-
- if (add_colon)
- FcStrBufChar (buf, ':');
- if (add_elt_name)
- {
- FcStrBufString (buf, c->word);
- FcStrBufChar (buf, '=');
- }
-
- l = FcPatternEltValues(e);
- FcNameUnparseValueList (buf, l, '\0');
- }
-
- return FcTrue;
-}
-
static FcBool
interpret_filter (FcFormatContext *c,
FcPattern *pat,
pass = pass &&
(negate ^
- (FcResultMatch == FcPatternGet (pat,
- (const char *) c->word,
- 0, &v)));
+ (FcResultMatch ==
+ FcPatternGet (pat, (const char *) c->word, 0, &v)));
}
while (consume_char (c, ','));
return FcTrue;
}
-static FcChar8 *
-convert (FcFormatContext *c,
- const FcChar8 *str)
+static FcBool
+interpret_count (FcFormatContext *c,
+ FcPattern *pat,
+ FcStrBuf *buf)
{
+ int count;
+ FcPatternElt *e;
+ FcChar8 buf_static[64];
+
+ if (!expect_char (c, '#'))
+ return FcFalse;
+
if (!read_word (c))
- return NULL;
- else if (0 == strcmp ((const char *) c->word, "downcase"))
- return FcStrDowncase (str);
- else if (0 == strcmp ((const char *) c->word, "basename"))
- return FcStrBasename (str);
- else if (0 == strcmp ((const char *) c->word, "dirname"))
- return FcStrDirname (str);
+ return FcFalse;
- message ("unknown converter \"%s\"",
- c->word);
- return NULL;
+ count = 0;
+ e = FcPatternObjectFindElt (pat,
+ FcObjectFromName ((const char *) c->word));
+ if (e)
+ {
+ FcValueListPtr l;
+ count++;
+ for (l = FcPatternEltValues(e);
+ l->next;
+ l = l->next)
+ count++;
+ }
+
+ snprintf ((char *) buf_static, sizeof (buf_static), "%d", count);
+ FcStrBufString (buf, buf_static);
+
+ return FcTrue;
}
static FcBool
-maybe_interpret_converts (FcFormatContext *c,
- FcStrBuf *buf,
- int start)
+interpret_array (FcFormatContext *c,
+ FcPattern *pat,
+ FcStrBuf *buf)
{
- while (consume_char (c, '|'))
+ FcObjectSet *os;
+ FcPattern *subpat;
+ const FcChar8 *format_save;
+ int idx;
+ FcBool ret, done;
+ FcStrList *lang_strs;
+
+ if (!expect_char (c, '[') ||
+ !expect_char (c, ']'))
+ return FcFalse;
+
+ os = FcObjectSetCreate ();
+ if (!os)
+ return FcFalse;
+
+ ret = FcTrue;
+
+ do
{
- const FcChar8 *str;
- FcChar8 *new_str;
+ if (!read_word (c) ||
+ !FcObjectSetAdd (os, (const char *) c->word))
+ {
+ FcObjectSetDestroy (os);
+ return FcFalse;
+ }
+ }
+ while (consume_char (c, ','));
- /* nul-terminate the buffer */
- FcStrBufChar (buf, '\0');
- if (buf->failed)
+ /* If we have one element and it's of type FcLangSet, we want
+ * to enumerate the languages in it. */
+ lang_strs = NULL;
+ if (os->nobject == 1)
+ {
+ FcLangSet *langset;
+ if (FcResultMatch ==
+ FcPatternGetLangSet (pat, os->objects[0], idx, &langset))
+ {
+ FcStrSet *ss;
+ if (!(ss = FcLangSetGetLangs (langset)) ||
+ !(lang_strs = FcStrListCreate (ss)))
+ goto bail0;
+ }
+ }
+
+ subpat = FcPatternDuplicate (pat);
+ if (!subpat)
+ goto bail0;
+
+ format_save = c->format;
+ idx = 0;
+ do
+ {
+ int i;
+
+ done = FcTrue;
+
+ if (lang_strs)
+ {
+ FcChar8 *lang;
+
+ FcPatternDel (subpat, os->objects[0]);
+ if ((lang = FcStrListNext (lang_strs)))
+ {
+ FcPatternAddString (subpat, os->objects[0], lang);
+ done = FcFalse;
+ }
+ }
+ else
+ {
+ for (i = 0; i < os->nobject; i++)
+ {
+ FcValue v;
+
+ /* XXX this can be optimized by accessing valuelist linked lists
+ * directly and remembering where we were. Most (all) value lists
+ * in normal uses are pretty short though (language tags are
+ * stored as a LangSet, not separate values.). */
+ FcPatternDel (subpat, os->objects[i]);
+ if (FcResultMatch ==
+ FcPatternGet (pat, os->objects[i], idx, &v))
+ {
+ FcPatternAdd (subpat, os->objects[i], v, FcFalse);
+ done = FcFalse;
+ }
+ }
+ }
+
+ if (!done)
+ {
+ c->format = format_save;
+ ret = interpret_subexpr (c, subpat, buf);
+ if (!ret)
+ goto bail;
+ }
+
+ idx++;
+ } while (!done);
+
+ if (c->format == format_save)
+ skip_subexpr (c);
+
+bail:
+ FcPatternDestroy (subpat);
+bail0:
+ if (lang_strs)
+ FcStrListDone (lang_strs);
+ FcObjectSetDestroy (os);
+
+ return ret;
+}
+
+static FcBool
+interpret_simple (FcFormatContext *c,
+ FcPattern *pat,
+ FcStrBuf *buf)
+{
+ FcPatternElt *e;
+ FcBool add_colon = FcFalse;
+ FcBool add_elt_name = FcFalse;
+ int idx;
+ FcChar8 *else_string;
+
+ if (consume_char (c, ':'))
+ add_colon = FcTrue;
+
+ if (!read_word (c))
+ return FcFalse;
+
+ idx = -1;
+ if (consume_char (c, '['))
+ {
+ idx = strtol ((const char *) c->format, (char **) &c->format, 10);
+ if (idx < 0)
+ {
+ message ("expected non-negative number at %d",
+ c->format-1 - c->format_orig + 1);
return FcFalse;
- str = buf->buf + start;
+ }
+ if (!expect_char (c, ']'))
+ return FcFalse;
+ }
- if (!(new_str = convert (c, str)))
+ if (consume_char (c, '='))
+ add_elt_name = FcTrue;
+
+ /* modifiers */
+ else_string = NULL;
+ if (consume_char (c, ':'))
+ {
+ FcChar8 *orig;
+ /* divert the c->word for now */
+ orig = c->word;
+ c->word = c->word + strlen ((const char *) c->word) + 1;
+ /* for now we just support 'default value' */
+ if (!expect_char (c, '-') ||
+ !read_chars (c, '\0'))
+ {
+ c->word = orig;
return FcFalse;
+ }
+ else_string = c->word;
+ c->word = orig;
+ }
+
+ e = FcPatternObjectFindElt (pat,
+ FcObjectFromName ((const char *) c->word));
+ if (e)
+ {
+ FcValueListPtr l;
+
+ if (add_colon)
+ FcStrBufChar (buf, ':');
+ if (add_elt_name)
+ {
+ FcStrBufString (buf, c->word);
+ FcStrBufChar (buf, '=');
+ }
- /* replace in the buffer */
- buf->len = start;
- FcStrBufString (buf, new_str);
- free (new_str);
+ l = FcPatternEltValues(e);
+
+ if (idx != -1)
+ {
+ while (l && idx > 0)
+ {
+ l = FcValueListNext(l);
+ idx--;
+ }
+ if (l && idx == 0)
+ {
+ if (!FcNameUnparseValue (buf, &l->value, '\0'))
+ return FcFalse;
+ }
+ else goto notfound;
+ }
+ else
+ {
+ FcNameUnparseValueList (buf, l, '\0');
+ }
+ }
+ else
+notfound:
+ {
+ if (else_string)
+ printf ("%s", else_string);
}
return FcTrue;
}
+static FcBool
+cescape (FcFormatContext *c,
+ const FcChar8 *str,
+ FcStrBuf *buf)
+{
+ while(*str)
+ {
+ switch (*str)
+ {
+ case '\\':
+ case '"':
+ FcStrBufChar (buf, '\\');
+ break;
+ }
+ FcStrBufChar (buf, *str++);
+ }
+ return FcTrue;
+}
+
+static FcBool
+shescape (FcFormatContext *c,
+ const FcChar8 *str,
+ FcStrBuf *buf)
+{
+ FcStrBufChar (buf, '\'');
+ while(*str)
+ {
+ if (*str == '\'')
+ FcStrBufString (buf, (const FcChar8 *) "'\\''");
+ else
+ FcStrBufChar (buf, *str);
+ str++;
+ }
+ FcStrBufChar (buf, '\'');
+ return FcTrue;
+}
+
+static FcBool
+xmlescape (FcFormatContext *c,
+ const FcChar8 *str,
+ FcStrBuf *buf)
+{
+ while(*str)
+ {
+ switch (*str)
+ {
+ case '&': FcStrBufString (buf, (const FcChar8 *) "&"); break;
+ case '<': FcStrBufString (buf, (const FcChar8 *) "<"); break;
+ case '>': FcStrBufString (buf, (const FcChar8 *) ">"); break;
+ default: FcStrBufChar (buf, *str); break;
+ }
+ str++;
+ }
+ return FcTrue;
+}
+
+static FcBool
+delete_chars (FcFormatContext *c,
+ const FcChar8 *str,
+ FcStrBuf *buf)
+{
+ /* XXX not UTF-8 aware */
+
+ if (!expect_char (c, '(') ||
+ !read_chars (c, ')') ||
+ !expect_char (c, ')'))
+ return FcFalse;
+
+ while(*str)
+ {
+ FcChar8 *p;
+
+ p = (FcChar8 *) strpbrk ((const char *) str, (const char *) c->word);
+ if (p)
+ {
+ FcStrBufData (buf, str, p - str);
+ str = p + 1;
+ }
+ else
+ {
+ FcStrBufString (buf, str);
+ break;
+ }
+
+ }
+
+ return FcTrue;
+}
+
+static FcBool
+escape_chars (FcFormatContext *c,
+ const FcChar8 *str,
+ FcStrBuf *buf)
+{
+ /* XXX not UTF-8 aware */
+
+ if (!expect_char (c, '(') ||
+ !read_chars (c, ')') ||
+ !expect_char (c, ')'))
+ return FcFalse;
+
+ while(*str)
+ {
+ FcChar8 *p;
+
+ p = (FcChar8 *) strpbrk ((const char *) str, (const char *) c->word);
+ if (p)
+ {
+ FcStrBufData (buf, str, p - str);
+ FcStrBufChar (buf, c->word[0]);
+ FcStrBufChar (buf, *p);
+ str = p + 1;
+ }
+ else
+ {
+ FcStrBufString (buf, str);
+ break;
+ }
+
+ }
+
+ return FcTrue;
+}
+
+static FcBool
+translate_chars (FcFormatContext *c,
+ const FcChar8 *str,
+ FcStrBuf *buf)
+{
+ char *from, *to, repeat;
+ int from_len, to_len;
+
+ /* XXX not UTF-8 aware */
+
+ if (!expect_char (c, '(') ||
+ !read_chars (c, ',') ||
+ !expect_char (c, ','))
+ return FcFalse;
+
+ from = (char *) c->word;
+ from_len = strlen (from);
+ to = from + from_len + 1;
+
+ /* hack: we temporarily divert c->word */
+ c->word = (FcChar8 *) to;
+ if (!read_chars (c, ')'))
+ {
+ c->word = (FcChar8 *) from;
+ return FcFalse;
+ }
+ c->word = (FcChar8 *) from;
+
+ to_len = strlen (to);
+ repeat = to[to_len - 1];
+
+ if (!expect_char (c, ')'))
+ return FcFalse;
+
+ while(*str)
+ {
+ FcChar8 *p;
+
+ p = (FcChar8 *) strpbrk ((const char *) str, (const char *) from);
+ if (p)
+ {
+ int i;
+ FcStrBufData (buf, str, p - str);
+ i = strchr (from, *p) - from;
+ FcStrBufChar (buf, i < to_len ? to[i] : repeat);
+ str = p + 1;
+ }
+ else
+ {
+ FcStrBufString (buf, str);
+ break;
+ }
+
+ }
+
+ return FcTrue;
+}
+
+static FcBool
+interpret_convert (FcFormatContext *c,
+ FcStrBuf *buf,
+ int start)
+{
+ const FcChar8 *str;
+ FcChar8 *new_str;
+ FcStrBuf new_buf;
+ FcChar8 buf_static[8192];
+ FcBool ret;
+
+ if (!expect_char (c, '|') ||
+ !read_word (c))
+ return FcFalse;
+
+ /* prepare the buffer */
+ FcStrBufChar (buf, '\0');
+ if (buf->failed)
+ return FcFalse;
+ str = buf->buf + start;
+ buf->len = start;
+
+ /* try simple converters first */
+ if (0) { }
+#define CONVERTER(name, func) \
+ else if (0 == strcmp ((const char *) c->word, name))\
+ do { new_str = func (str); ret = FcTrue; } while (0)
+ CONVERTER ("downcase", FcStrDowncase);
+ CONVERTER ("basename", FcStrBasename);
+ CONVERTER ("dirname", FcStrDirname);
+#undef CONVERTER
+ else
+ ret = FcFalse;
+
+ if (ret)
+ {
+ if (new_str)
+ {
+ FcStrBufString (buf, new_str);
+ free (new_str);
+ return FcTrue;
+ }
+ else
+ return FcFalse;
+ }
+
+ FcStrBufInit (&new_buf, buf_static, sizeof (buf_static));
+
+ /* now try our custom converters */
+ if (0) { }
+#define CONVERTER(name, func) \
+ else if (0 == strcmp ((const char *) c->word, name))\
+ ret = func (c, str, &new_buf)
+ CONVERTER ("cescape", cescape);
+ CONVERTER ("shescape", shescape);
+ CONVERTER ("xmlescape", xmlescape);
+ CONVERTER ("delete", delete_chars);
+ CONVERTER ("escape", escape_chars);
+ CONVERTER ("translate", translate_chars);
+#undef CONVERTER
+ else
+ ret = FcFalse;
+
+ if (ret)
+ {
+ FcStrBufChar (&new_buf, '\0');
+ FcStrBufString (buf, new_buf.buf);
+ }
+ else
+ message ("unknown converter \"%s\"",
+ c->word);
+
+ FcStrBufDestroy (&new_buf);
+
+ return ret;
+}
+
+static FcBool
+maybe_interpret_converts (FcFormatContext *c,
+ FcStrBuf *buf,
+ int start)
+{
+ while (*c->format == '|')
+ if (!interpret_convert (c, buf, start))
+ return FcFalse;
+
+ return FcTrue;
+}
+
static FcBool
align_to_width (FcStrBuf *buf,
int start,
start = buf->len;
switch (*c->format) {
+ case '=': ret = interpret_builtin (c, pat, buf); break;
case '{': ret = interpret_subexpr (c, pat, buf); break;
case '+': ret = interpret_filter (c, pat, buf); break;
case '-': ret = interpret_delete (c, pat, buf); break;
case '?': ret = interpret_cond (c, pat, buf); break;
+ case '#': ret = interpret_count (c, pat, buf); break;
+ case '[': ret = interpret_array (c, pat, buf); break;
default: ret = interpret_simple (c, pat, buf); break;
}
expect_char (c, '}');
}
-static char escaped_char(const char ch)
-{
- switch (ch) {
- case 'a': return '\a';
- case 'b': return '\b';
- case 'f': return '\f';
- case 'n': return '\n';
- case 'r': return '\r';
- case 't': return '\t';
- case 'v': return '\v';
- default: return ch;
- }
-}
-
static FcBool
interpret_expr (FcFormatContext *c,
FcPattern *pat,
return FcTrue;
}
-FcChar8 *
-FcPatternFormat (FcPattern *pat, const FcChar8 *format)
+static FcBool
+FcPatternFormatToBuf (FcPattern *pat,
+ const FcChar8 *format,
+ FcStrBuf *buf)
{
- FcStrBuf buf;
FcFormatContext c;
- FcBool ret;
+ FcChar8 word_static[1024];
+ FcBool ret;
- FcStrBufInit (&buf, 0, 0);
- if (!FcFormatContextInit (&c, format))
- return NULL;
+ if (!FcFormatContextInit (&c, format, word_static, sizeof (word_static)))
+ return FcFalse;
- ret = interpret_expr (&c, pat, &buf, '\0');
- if (buf.failed)
- ret = FcFalse;
+ ret = interpret_expr (&c, pat, buf, '\0');
FcFormatContextDone (&c);
+
+ return ret;
+}
+
+FcChar8 *
+FcPatternFormat (FcPattern *pat,
+ const FcChar8 *format)
+{
+ FcStrBuf buf;
+ FcChar8 buf_static[8192 - 1024];
+ FcBool ret;
+
+ FcStrBufInit (&buf, buf_static, sizeof (buf_static));
+
+ ret = FcPatternFormatToBuf (pat, format, &buf);
+
if (ret)
return FcStrBufDone (&buf);
else