| Commit message (Collapse) | Author | Age | Files | Lines | |
|---|---|---|---|---|---|
| * | Squash another warning | John-Mark Bell | 2018-01-22 | 1 | -1/+1 |
| | | |||||
| * | Squash warning | John-Mark Bell | 2018-01-22 | 1 | -1/+1 |
| | | |||||
| * | "Old" FontManager: improve Encoding file parser. | John-Mark Bell | 2018-01-22 | 1 | -12/+55 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 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. | ||||
| * | Fix typo | John-Mark Bell | 2018-01-22 | 1 | -1/+1 |
| | | |||||
| * | "Old" FontManager: log character being scanned too. | John-Mark Bell | 2018-01-22 | 1 | -2/+2 |
| | | |||||
| * | Fix typo in log message | John-Mark Bell | 2018-01-21 | 1 | -1/+1 |
| | | |||||
| * | Old FontManager: ignore fonts with no mappings | John-Mark Bell | 2018-01-21 | 1 | -0/+7 |
| | | | | | | | | | | | | Ordinarily, this shouldn't happen but, if it does, we don't want to pollute the cache with a blank charset for the font (as we'd prefer to rescan the font next time around in case whatever caused it to have no mappings got fixed in the interim). Existing caches containing blank charsets for such fonts will remain untouched -- this is all best effort, so forcing a rescan by bumping the cache version is not worth doing here. | ||||
| * | Accept FontEncodingNotFound for symbol fonts, too. | John-Mark Bell | 2018-01-21 | 1 | -2/+4 |
| | | |||||
| * | Tolerate FontEncodingNotFound from "old" FM. | John-Mark Bell | 2018-01-21 | 1 | -5/+9 |
| | | | | | Additionally, log the FontManager version. | ||||
| * | Log in release builds; enhance init logging. | John-Mark Bell | 2018-01-21 | 2 | -8/+40 |
| | | |||||
| * | Add times to debug output. | Michael Drake | 2012-11-06 | 1 | -3/+16 |
| | | |||||
| * | put makeglyphs tool back | Vincent Sanders | 2012-07-17 | 1 | -1/+1 |
| | | |||||
| * | Update to new NSBUILD infrastructure | Daniel Silverstone | 2012-06-29 | 1 | -1/+1 |
| | | | | | svn path=/trunk/rufl/; revision=14012 | ||||
| * | Only build strfuncs.c if Norcroft is the toolchainreleases/0.0.1 | John Mark Bell | 2011-01-02 | 1 | -1/+5 |
| | | | | | svn path=/trunk/rufl/; revision=11185 | ||||
| * | Port to core buildsystem. | John Mark Bell | 2010-01-06 | 14 | -0/+4879 |
| The python module (and associated make runes) need some love (as does non-GCC building with the core buildsystem in general) svn path=/trunk/rufl/; revision=9792 | |||||
 |
index : librufl.git | |
| RISC OS Unicode Font Library | NetSurf automatic user |
| summaryrefslogtreecommitdiff |