"Old" FontManager: improve Encoding file parser.

In a non-Unicode world, a (non-Base) encoding may define glyphs for
up to 256 character codes. Ensure that at most 256 Encoding file
entries are used (as, otherwise, the character code will overflow).

In particular, if symbol fonts created for the Unicode Font Manager
(which does not have a 256 character limit for an encoding) are
installed on a non-Unicode-capable system, only the first 256
glyphs in the font are accessible although the Encoding file may
have more than 256 entries.

Note, however, that the first 32 character codes will never be used
as they are considered control codes. Thus, at most 224 usable
characters may be defined.

A further wrinkle is that glyph names may map to multiple Unicode
codepoints, thus consuming multiple slots in the unicode map (which
itself has a fixed size of 256 entries). Thus, it is technically
possible for the unicode map to further limit the number of usable
characters in a font to fewer than 224.

However, unless the font is particularly baroque, this isn't a
problem in the real world, because there are only 12 glyph names
which map to more than one Unicode codepoint (they map to 2, each,
for a total of 24 unicode map entries, if they're all present).
Thus, to run out of space in the unicode map, you'd need a font
which defines at least 4 of those glyphs twice (and defines the
others once, and also defines known glyphs for every other
character code).

Fixes #2577.

2 files changed, 62 insertions, 12 deletions

diff --git a/src/rufl_init.c b/src/rufl_init.c
index 9c80b0a..8cd0358 100644
--- a/src/rufl_init.c
+++ b/src/rufl_init.c
@@ -1011,6 +1011,12 @@ rufl_code rufl_init_scan_font_in_encoding(const char *font_name,
 rufl_code rufl_init_read_encoding(font_f font,
 		struct rufl_unicode_map *umap)
 {
+	enum {
+		STATE_START,
+		STATE_COMMENT,
+		STATE_COLLECT,
+	} state = STATE_START;
+	bool emit = false;
 	unsigned int u = 0;
 	unsigned int i = 0;
 	int c;
@@ -1035,18 +1041,50 @@ rufl_code rufl_init_read_encoding(font_f font,
 	if (!fp)
 		return rufl_IO_ERROR;
 
-	while (!feof(fp) && u != 256) {
+	while (!feof(fp) && i < 256 && u < 256) {
 		c = fgetc(fp);
-		if (c == '%') {
-			/* comment line */
-			fgets(s, sizeof s, fp);
-		} else if (c == '/') {
-			/* character definition */
-			if (i++ < 32)
-				continue;
-			n = fscanf(fp, "%100s", s);
-			if (n != 1)
-				break;
+
+		if (state == STATE_START) {
+			if (c == '/') {
+				n = 0;
+				state = STATE_COLLECT;
+			} else if (c <= 0x20) {
+				/* Consume C0 and space */
+			} else {
+				/* Comment, or garbage */
+				state = STATE_COMMENT;
+			}
+		} else if (state == STATE_COMMENT) {
+			/* Consume until the next C0 */
+			if (c < 0x20) {
+				state = STATE_START;
+			}
+		} else {
+			if ((c >= '0' && c <= '9') ||
+					(c >= 'a' && c <= 'z') ||
+					(c >= 'A' && c <= 'Z') ||
+					(c == '.') || (c == '_')) {
+				/* Printable: append */
+				s[n++] = c;
+				if (n >= sizeof(s)) {
+					/* Too long: garbage */
+					state = STATE_COMMENT;
+				}
+			} else if (c > 0x20) {
+				/* Garbage */
+				state = STATE_COMMENT;
+			} else {
+				/* C0 or space: done */
+				s[n] = '\0';
+				if (n != 0) {
+					emit = true;
+				}
+				state = STATE_START;
+			}
+		}
+
+		/* Ignore first 32 character codes (these are control chars) */
+		if (emit && i > 31 && i < 256 && u < 256) {
 			entry = bsearch(s, rufl_glyph_map,
 					rufl_glyph_map_size,
 					sizeof rufl_glyph_map[0],
@@ -1059,13 +1097,18 @@ rufl_code rufl_init_read_encoding(font_f font,
 				for (; strcmp(s, entry->glyph_name) == 0;
 						entry++) {
 					umap->map[u].u = entry->u;
-					umap->map[u].c = i - 1;
+					umap->map[u].c = i;
 					u++;
 					if (u == 256)
 						break;
 				}
 			}
 		}
+
+		if (emit) {
+			i++;
+			emit = false;
+		}
 	}
 
 	if (fclose(fp) == EOF)
diff --git a/test/data/Encoding b/test/data/Encoding
new file mode 100644
index 0000000..e5b633e
--- /dev/null
+++ b/test/data/Encoding
@@ -0,0 +1,7 @@
+% Encoding file for font 'Allerta'
+/
+/x
+/w
+   % xyz abc/y
+  /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /space /exclam /quotedbl /.notdef /dollar /percent /ampersand /quotesingle /parenleft /parenright /asterisk /plus /comma /hyphen /period /slash /zero /one /two /three /four /five /six /seven /eight /nine /colon /semicolon /less /equal /greater /question /at /A /B /C /D /E /F /G /H /I /J /K /L /M /N /O /P /Q /R /S /T /U /V /W /X /Y /Z /bracketleft /backslash /bracketright /.notdef /.notdef /grave /a /b /c /d /e /f /g /h /i /j /k /l /m /n /o /p /q /r /s /t /u /v /w /x /y /z /.notdef /bar /.notdef /.notdef /.notdef /Euro /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /quoteleft /quoteright /guilsinglleft /guilsinglright /quotedblleft /quotedblright /.notdef /endash /emdash /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /space /exclamdown /cent /sterling /currency /yen /.notdef /.notdef /dieresis /.notdef /.notdef /guillemotleft /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /.notdef /acute /.notdef /.notdef /.notdef /cedilla /.notdef /.notdef /guillemotright /.notdef /.notdef /.notdef /questiondown /Agrave /Aacute /Acircumflex /Atilde /Adieresis /Aring /.notdef /Ccedilla /Egrave /Eacute /Ecircumflex /Edieresis /Igrave /Iacute /Icircumflex /Idieresis /Eth /Ntilde /Ograve /Oacute /Ocircumflex /Otilde /Odieresis /multiply /.notdef /Ugrave /Uacute /Ucircumflex /Udieresis /Yacute /Thorn /germandbls /agrave /aacute /acircumflex /atilde /adieresis /aring /.notdef /ccedilla /egrave /eacute /ecircumflex /edieresis /igrave /iacute /icircumflex /idieresis /eth /ntilde /ograve /oacute /ocircumflex /otilde /odieresis /divide /.notdef /ugrave /uacute /ucircumflex /udieresis /yacute /thorn /ydieresis /dotlessi /Lslash /lslash /Scaron /scaron /Ydieresis /Zcaron /zcaron /circumflex /caron /breve /dotaccent /ring /ogonek /tilde /hungarumlaut 
+