b9a176e268
[subset] speedup cmap4 subsetting for large codepoint counts. ( #3178 )
...
glyphIdArray generation implementation was O(n^2). Refactored to use a hashmap to reduce complexity. After the change subset time for a 22k codepoint subset went from 7s to 0.7s.
2021-08-29 10:33:12 -06:00
57d6bf8251
[subset] in sets union correctly size the pointer array.
2021-08-27 10:59:28 -07:00
102666dab7
[subset] add test for hb_subset_input_set (...).
2021-08-27 10:36:42 -07:00
7194c2a2a3
[subset] use anonymous union to hold the input sets instead of a map.
2021-08-27 10:36:42 -07:00
05204d7586
[subset] implement hb_subset_input_set (...).
...
Switch to storing the sets keyed by enum internally.
2021-08-27 10:36:36 -07:00
50193262f0
[subset] reorder set enums.
2021-08-27 10:34:59 -07:00
3282f540a9
[subset] restore legacy set api methods.
2021-08-27 10:34:58 -07:00
1457c1f080
[subset] set enum name changes.
2021-08-27 10:34:58 -07:00
1b5e159301
[subset] Remove HB_SUBSET_SETS_COUNT and flags comment.
2021-08-27 10:34:58 -07:00
4c389fd9e5
[subset] add unicode and glyph set enums.
2021-08-27 10:34:58 -07:00
e3e1ae8766
[subset] proposed update to subset set apis.
2021-08-27 10:34:58 -07:00
9db494f517
[subset] Update remaining subset input sets to use inversion for all. ( #3172 )
...
* [subset] Update remaining subset input sets to use inversion for all.
* [subset] Update flag help for remaining sets to mention -=.
2021-08-26 10:51:38 -06:00
99356ea5f0
Merge pull request #3150 from harfbuzz/item-context-tests
...
[test] Add tests for item context
2021-08-25 17:19:06 -06:00
1620698bd5
[subset-input] Minor format
2021-08-25 16:26:58 -06:00
cd07070e41
[subset] Move plan unicodes and gids population to end of _populate_unicodes...
2021-08-25 14:42:00 -07:00
4b5a81f13c
[buffer] Fix hb_buffer_append() pre/post-context logic
...
Part of https://github.com/harfbuzz/harfbuzz/pull/3150
2021-08-25 15:20:54 -06:00
8f4f47df7c
[subset] use inverted set for all unicodes.
...
Modify the code that handles input->unicodes to be safe with possibly inverted sets. Also adds --unicodes-= and --unicodes+= flags.
2021-08-25 13:34:05 -07:00
e9e6d66cd6
[subset] Minor internal parameter rename
2021-08-24 23:22:49 -06:00
59deb754b3
[subset] rename _collect_subset_layout to _collect_layout_indices.
...
Better describes what the function does.
2021-08-24 18:17:07 -06:00
f2441a4b65
[subset] Remove retain all layout features flag.
...
Instead use inverted sets to handle requesting all features. Modifies feature collection in subset plan to intersect the set of requested features against the features in the font. This prevents iterating a fully filled feature tag set.
2021-08-24 18:17:07 -06:00
c90678cbc5
[set] One more try to fix things
...
That I don't let a "make check -j10" run finish before pushing things
out means either that we need to speed up our build / test process,
or I need a faster laptop...
2021-08-24 11:11:12 -06:00
357976963e
[set] One more try at fixing clear() on empty set
2021-08-24 10:59:31 -06:00
fb07f8f876
During subset input creation check for set alloc failures and fail if encountered.
2021-08-24 10:59:14 -06:00
b5177d2147
[set] Fix seam exposed by previous commit
2021-08-24 10:43:37 -06:00
f0de40dd61
Whitespace
2021-08-24 10:35:33 -06:00
8cdbea5580
[ot-shape-normalize] Move buffer out of hb_ot_shape_normalize_context_t
2021-08-24 10:35:33 -06:00
10da9fd20d
[ot-map] Whitespace
2021-08-24 10:35:33 -06:00
d3e09bf465
[set] Make all operators null-safe again
...
Changed my mind.
Also for hb_map_clear().
Part of https://github.com/harfbuzz/harfbuzz/pull/3162
2021-08-24 10:31:49 -06:00
430224b1a0
[buffer] Handle pre/post-context in buffer_append
...
Fixes https://github.com/harfbuzz/harfbuzz/issues/1843
2021-08-24 02:11:17 +02:00
33c82c7dad
[ot-shape] Fix unused-var error when building without AAT
...
Fixes https://github.com/harfbuzz/harfbuzzjs/issues/45
2021-08-22 22:31:39 -06:00
34e0b28fae
[subset] make glyph-names option match fonttools behavior
...
Don't encode psNames that are contained in standard glyph names set
2021-08-20 01:30:23 -06:00
829b0f3353
Merge pull request #3154 from harfbuzz/set-invert
...
Resurrect hb_set_invert()
2021-08-20 01:28:15 -06:00
84c2a30214
[bit-set-invertible] Second try fixing previous()
2021-08-19 16:33:48 -06:00
8f88747a5e
[bit-set-invertible] Fix previous() iteration
2021-08-19 16:31:01 -06:00
1d832693e1
[set] Protect against immutible null set with invertible addition
2021-08-19 16:02:30 -06:00
dc800ffd6c
[bit-set] Remove TODO items we don't intend to finish
2021-08-19 15:41:27 -06:00
92908c122b
[bit-set-invertible] Remove extra check
...
We don't expect immutable / null set in this code.
2021-08-19 15:28:26 -06:00
1babe80ed2
[bit-set-invertible] Fix last remaining corner-case
2021-08-19 15:18:32 -06:00
e8911d137c
[bit-set-invertible] Implement a couple other missing corner cases
...
Also, in bit-set, don't compute population in is_equal/is_subset()
if we don't have it computed already.
2021-08-19 15:00:33 -06:00
2579dc648f
[bit-set-invertible] Move code around
2021-08-19 14:48:27 -06:00
2dfc104236
[bit-set-invertible] Implement is_empty() for inverted case as well
2021-08-19 14:47:33 -06:00
ec4812a7d4
[bit-page] Remove unused inverted code
2021-08-19 13:32:44 -06:00
87885e6f02
[bit-set-invertible] Fix next/previous() logic
2021-08-19 13:26:25 -06:00
ee4e331a6e
[bit-set] Remove unused get_min(inverted)
2021-08-19 13:24:15 -06:00
eec1a25e71
[bit-set-invertible] Rewrite get_min/max() in terms of next/previous()
2021-08-19 13:24:00 -06:00
f09d5ed0f5
[algs] Remove hb_bitwise_non()
2021-08-19 13:06:10 -06:00
f4fd7baf7e
[bit-set-invertible] Fix subtract() logic
2021-08-19 13:05:12 -06:00
f317d8e426
[bit-set-invertible] Fix intersect() logic
2021-08-19 13:01:07 -06:00
b21038d91d
[bit-set-invertible] Rewrite next/previous_range() in terms of s.next/previous()
2021-08-19 11:24:17 -06:00
eb98bc1e32
[bit-set-invertible] Rewrite next/previous() using s.next/previous_range()
2021-08-19 11:19:24 -06:00
248ad3bce5
[bit-set-invertible] Implement next/previous
...
This makes invertible set functionality complete.
2021-08-19 10:17:51 -06:00
c27f5b1288
[bit-set-invertible] Plug in next_range()/previous_range()
2021-08-19 10:01:24 -06:00
f6aa37159b
[bit-set-invertible] Fix process logic for inverted
...
[# On branch set-invert
2021-08-19 01:12:25 -06:00
060de189ec
[bit-page] Fix sanitizer error
...
This essentially reverts 9449cfeefd
2021-08-19 00:58:34 -06:00
b119b48079
[bit-set-invertible] Add unlikely() around inverted checks
2021-08-19 00:51:24 -06:00
b94f24ec79
[bit-set-invertible] Implement set algebra
2021-08-19 00:48:00 -06:00
6afefe1dc3
[algs] Remove unnecessary struct tag names
2021-08-19 00:27:03 -06:00
8aa92ff8f0
[bit-set-invertible] Implement get_min/max
2021-08-18 23:01:06 -06:00
18f50275ed
[bit-set] Restructure get_min/max() in prep for adding inverted
2021-08-18 22:08:06 -06:00
669b97d949
[bit-set-invertible] Implement iterator
2021-08-18 21:39:04 -06:00
c66894d7c9
[bit-set-invertible] Write a subset branch as dagger
2021-08-18 21:24:29 -06:00
c39d2f7a76
[bit-set-invertible] Implement add_array / add_sorted_array / intersects
2021-08-18 21:20:54 -06:00
0efa614c04
[bit-set] Add del_array/del_sorted_array()
2021-08-18 21:18:32 -06:00
48ad9eef1e
[bit-set] Merge page_for_insert() into page_for()
2021-08-18 21:05:21 -06:00
280366ba6a
Add TODO to update NN offsets to allow nulls.
2021-08-18 16:52:45 -06:00
dc31920bbe
Don't serialize null offsets in CPAL.
...
Fixes https://oss-fuzz.com/testcase-detail/5443213648330752
2021-08-18 16:52:45 -06:00
80ff72f39d
[doc] Misc fixes
2021-08-18 13:16:51 +02:00
9aa6f8a93f
2.9.0
2021-08-18 09:32:04 +02:00
f0c3804fa2
[set] Add hb_bit_set_invertible_t and resurrect hb_set_invert()
...
Implementation is NOT complete yet.
2021-08-17 19:53:43 -06:00
b5cdbdc030
[set] Turn hb_set_t into instance of template taking implementation t
2021-08-17 10:53:08 -06:00
fad452bffb
[set] Move main functionality into hb_bit_set_t
...
To add inversion on top in hb_invertible_set_t and use that as hb_set_t.
2021-08-17 10:47:18 -06:00
9cc4da962f
[object] Remove hb_object_is_inert()
2021-08-16 20:17:53 -06:00
2d5ef05d67
Use !g_object_is_valid() instead of g_object_is_inert()
2021-08-16 20:13:58 -06:00
f6d6eff6a2
[object] Remove unused HB_REFERENCE_COUNT_INIT
2021-08-16 19:41:44 -06:00
f245dc4db8
[set] Whitespace
2021-08-15 18:59:58 -06:00
9b390f8c40
[set] Move page_t into hb-bit-page.hh
2021-08-15 18:59:58 -06:00
0c3e02ee2d
[set] Add "inverted" to page_t::is_empty()
2021-08-15 18:59:58 -06:00
9bd64fa077
[set] Add "inverted" to page_t::next/prev()
2021-08-15 18:59:58 -06:00
c88e7ec935
[set] Add page_t::elt_maybe_invert()
2021-08-15 18:59:58 -06:00
9449cfeefd
[set] Simplify page_t::next/prev()
2021-08-15 18:59:58 -06:00
4394ee1f1d
[set] Add inverted to page_t::get_min/max()
2021-08-15 18:59:58 -06:00
0dcd9b15d9
[set] Add page_t::set_range()
2021-08-15 18:59:58 -06:00
cb273fd17c
[set] Add page_t::set()
2021-08-15 18:59:58 -06:00
32bbf53d9c
[set] Minor rewrite one cmp() in term of other
2021-08-15 18:59:58 -06:00
cb05c363d7
[subset] add some basic tests for the legacy subset methods.
2021-08-14 10:37:02 -06:00
d18aad7ec5
[subset] restore old subsetting api methods.
...
These will be left in place until the 3.0.0 release.
2021-08-14 10:37:02 -06:00
c0f3af91b8
[subset] speed up add_gid_and_children and adjust op limit.
...
Fix for fuzzer timeout: https://oss-fuzz.com/testcase-detail/5001604901240832 .
- Operation limit is per glyph, so 100,000 should still be far more than needed.
- Switches from for(...) to while(...) loop for iteration. for(...) calls it.end() which in this case triggers a complete iteration.
- Cache CompositeGlyph size in the iterator to avoid needing to recalculate it.
2021-08-12 14:50:42 -06:00
c08f1b8903
[map] fix incorrect population count in hash map.
...
If the same key was set twice the population was being incorrectly incremented.
2021-08-10 14:00:55 -06:00
e39c3bde7b
[subset] fix bug in parsing glyf flags.
...
the bytes of the flag stream were being treated as signed integers instead of unsigned as specified in the spec.
2021-08-10 12:57:07 -06:00
f5ba1fabce
Fix documentation of hb_subset_flags_t defaults
...
HB_SUBSET_FLAGS_DEFAULT is 0, and hb_subset_input_create_or_fail
initializes its flags to HB_SUBSET_FLAGS_DEFAULT, so currently all flags
are set to false by default. It has been decided that future flags will
also be false by default for simplicity (using NO_ in the name if
needed). Correct and simplify the documentation for hb_subset_flags_t to
reflect this.
2021-08-09 09:05:24 -06:00
8c0c217b5a
[subset] fail reference blob in face builder if allocation for table sorting fails.
...
Fixes https://oss-fuzz.com/testcase-detail/5041767803125760
2021-08-06 15:54:41 -06:00
e5bfd49ae5
[subset] don't allow table adds for tag == -1.
2021-08-05 17:49:37 -06:00
222b74f08f
[subset] don't leak memory for duplicate tables.
2021-08-05 17:49:37 -06:00
f7a9c3ea93
[open-file] Use hb_is_source_of instead of hb_is_iterator
2021-08-05 13:05:07 -06:00
fa90cb8a31
Whitespace again
2021-08-05 12:31:06 -06:00
a110a47eb5
Whitespace
2021-08-05 12:28:39 -06:00
2cb8c928d7
[face] Let hashmap derive invalid key/value types
...
Works now with the previous two fixes.
2021-08-05 12:27:22 -06:00
95c888e192
[meta] Define hb_int_min for pointer types as nullptr
2021-08-05 12:27:02 -06:00
407a112e7b
[meta] Make hb_is_signed/hb_is_unsigned work on all types
...
They were failing if type was non-scalar, eg. pointer.
2021-08-05 12:23:53 -06:00
baf2166a24
[meta] Use hb_true_type/hb_false_type more
2021-08-05 12:08:47 -06:00
Garret Rieger
Behdad Esfahbod
Khaled Hosny
Qunxin Liu
Ben Wagner