Fixbumps 2016 03#3
Merged
Merged
Conversation
mattleibow
pushed a commit
that referenced
this pull request
Jun 11, 2020
Add AOT support for libSkia
1 task
1 task
mattleibow
added a commit
that referenced
this pull request
Apr 10, 2026
- Gotcha #6 (BUILD.gn legacy flags): Replace 'comment these out' with a proper investigation flow: check what the flag removes, update the C API to use the replacement, only comment out as a tracked short-term bridge with a TODO - Gotcha #3 (custom patches): Add concrete steps for when declarations diverge — check for upstream replacement, consider moving custom methods into C API layer, never leave mismatched header/impl - Phase 5 error table: Flag removed enum values as C# breaking changes requiring [Obsolete] or documentation in Phase 8. Add legacy flag row pointing to gotcha #6 - Fix duplicate step 4/5 in Phase 2 (copy-paste from earlier edit) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
mattleibow
added a commit
that referenced
this pull request
Apr 26, 2026
SKTypeface Span methods now handle Skia's all-or-nothing semantics: if the buffer is too small, retry with a pooled buffer and copy what fits, returning the actual written count instead of the total. Tests: - Fix CanSetMultipleVariationsSimultaneously silent pass (#3) - Add undersized buffer tests for HarfBuzz and Skia Span methods (#4) - Tighten all assertions to exact values - Add InterVariable.ttf (multi-axis) for undersized buffer testing Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
mattleibow
added a commit
that referenced
this pull request
Apr 26, 2026
SKTypeface Span methods now handle Skia's all-or-nothing semantics: if the buffer is too small, retry with a pooled buffer and copy what fits, returning the actual written count instead of the total. Font.GetVariationCoordsNormalized(Span) also fixed to return written count instead of total length when buffer is undersized. Tests: - Fix CanSetMultipleVariationsSimultaneously silent pass (#3) - Add undersized buffer tests for HarfBuzz and Skia Span methods (#4) - Tighten all assertions to exact values - Add InterVariable.ttf (multi-axis) for undersized buffer testing Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
mattleibow
added a commit
that referenced
this pull request
Apr 26, 2026
SKTypeface Span methods now handle Skia's all-or-nothing semantics: if the buffer is too small, retry with a pooled buffer and copy what fits, returning the actual written count instead of the total. Font.GetVariationCoordsNormalized(Span) also fixed to return written count instead of total length when buffer is undersized. Tests: - Fix CanSetMultipleVariationsSimultaneously silent pass (#3) - Add undersized buffer tests for HarfBuzz and Skia Span methods (#4) - Tighten all assertions to exact values - Add InterVariable.ttf (multi-axis) for undersized buffer testing Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
mattleibow
added a commit
that referenced
this pull request
May 25, 2026
- glibc/Dockerfile: Extract only the 6 required packages from debootstrap cache instead of all cached .debs. This eliminates the 'Cannot create symlink' errors from base-files and removes the grep filter workaround. - loongarch64/Dockerfile: Replace HTML scraping with hardcoded package URLs and SHA-256 checksums for all 7 cross-packages. Reproducible and verifiable — no more runtime HTML parsing. - loongarch64/Dockerfile: Remove linker-script sed fixup entirely. Instead, export SYSROOT_ROOT=/sysroot so build.cake passes --sysroot=/sysroot to lld, which automatically prepends the sysroot to absolute paths in linker scripts. This is the same mechanism used by the glibc/debootstrap path. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
5 tasks
mattleibow
added a commit
that referenced
this pull request
Jul 6, 2026
…roster
Two Prepare-phase changes that make the Polish AI's product-focus and its
credits table reliable run-to-run instead of a per-run judgment call.
~~ Three-way PR classification (product / mixed / internal) ~~
The old tagger was binary and treated `native/` as shipped code, so every
build-config PR that touches `native/**/build.cake` (Docker migration, SDK
pins, ADO caching, devcontainer/Tizen CI splits) was tagged `[product]` and
leaked into the "Lifecycle & Internals" category as noise. But `native/` is
build configuration — the code that actually ships is `externals/skia/`.
Split the classifier by what a path means to a consumer:
* product — touches shipped code (binding/ externals/ source/); write it up.
* internal — touches none of those (CI, .agents skills, docs, tests,
samples, build/meta); drop into the one collapse line.
* mixed — touches only build config (native/); may change the shipped
binary via a compile define (a rasteriser flag, a delay-load
fix) or be pure infra. Polish guesses from the title/context
already in the raw-data block — it does NOT open the PR — and
surfaces it only when it plausibly changes what ships.
On real 4.148.0 data this moves the 6 build-config PRs from `[product]` to
`[mixed]` (so they drop) while a native compile-flag fix like the small-sigma
raster-blur regression lands as `[mixed]` and is surfaced by inspection rather
than hidden. Distribution over 272 recent PRs: 60 product / 10 mixed / 202
internal.
~~ Authoritative contributor roster ~~
The credits table was reconstructed by hand from the body prose, so any
contributor whose PRs were folded into a thematic bullet was silently dropped.
On 4.148.0 the table listed only 2 of 5 community members and omitted
@ramezgerges entirely — the single biggest external contributor (16 PRs:
variable fonts, singleton-lifecycle rework, default-typeface-to-managed, the
SKPath finalizer fix, several Skia bumps).
Prepare now emits a deterministic `contributors:` roster in the raw-data block
— every distinct non-maintainer, non-bot author (linkable login) with their PR
numbers, ordered by PR count — and Polish renders exactly one row per entry,
never omitting one. Bots (github-actions[bot], Copilot, dependabot, any
`*[bot]`) and @mattleibow are excluded at the source.
Spec-first: §4.4 principles #1 and #3 updated to define the three-way tag and
the roster-driven table; SKILL rules 1/6 + self-review, the review-checklist
(contributor-table-complete), and TEMPLATE.md brought in line.
Co-authored-by: Matthew Leibowitz <mattleibow@live.com>
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Fixing the submodule revision