Repository navigation
feat: add a library for reading compressed images - #817
Conversation
There was a problem hiding this comment.
Summary: This change adds the build recipe for a small library that decodes a compressed black-and-white image format used in scanned PDFs. The recipe, its source checksum, and the bookkeeping all look right; one bookkeeping record is missing the pull-request number that this project's own guide asks for, so reviewers later can trace where the recipe came from.
What I need from you: Add the PR number to the catalog record (one-line suggestion below).
style
-
config/factory-build-backlog.toml:354— the new[resolved]record is{ name = "jbig2dec-libs" }with nopr.docs/skills/factory-build-backlog.md:66and:81-83say a recipe-only record should be{ name = "<n>", pr = <n> }and thatprmust be populated before requesting review.--checkdoes not validate this field (per:82), so CI will not catch it.{ name = "jbig2dec-libs", pr = 817 },
docs-currency (out-of-scope, for later)
packages/ghostscript/ghostscript.spec:61-66— thesystem_jbig2dec_ijsbcond comment sayslibjbig2dec.so.0is "in neither Hummingbird nor this factory". Still true at this head (no hosted build or publication is claimed in the PR body), and the new skill section atdocs/skills/factory-build-backlog.md:907-909correctly notes Ghostscript keeps its bundled copy. Oncejbig2decactually publishes, that comment and bcond rationale will need revisiting; not this PR's job.
Verified for the record: packages/jbig2dec/sources:1 and config/upstream-sources.json:9 carry the same SHA-512 for the GitHub release tarball; the %if 0%{?fedora} >= 40 guard in packages/jbig2dec/jbig2dec.spec:4-8,62-66 resolves to current_version 0.20 in the factory's Fedora 44 mock root (.github/workflows/rebuild-rpms.yml:177,517), so the version-mismatch exit 1 does not trigger; trailing whitespace in the imported spec is excluded by .pre-commit-config.yaml (trailing-whitespace has exclude: ^packages/), matching the repo's keep-upstream-bytes rule.
No findings from: correctness, security, intent-alignment.
Confidence: 3/5 (needs attention) — 1 perspective requested changes
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Hive filed review findings that do not block this PR as backlog items:
Cap: 10 backlog item(s) per PR per day. |
|
@Karibhou this PR is back in your court ( Review asks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Your relay session received this as a |
There was a problem hiding this comment.
Summary: Re-review of the one-line follow-up since the last look. The backlog record now points at this pull request, which was the only change requested before. The remaining note is advisory: a comment in the PDF renderer's recipe says this library is not in the factory, which stops being true once this lands.
docs-currency (info, advisory — unchanged since previous review)
packages/ghostscript/ghostscript.spec:61-62 still says libjbig2dec.so.0 ... are in neither Hummingbird nor this factory. With packages/jbig2dec/jbig2dec.spec added here, the factory now carries a jbig2dec recipe (unpublished). Suggested wording so it does not go stale at publication:
# jbig2dec and ijs are built from the copies in the tarball. This factory now
# carries a jbig2dec recipe (not yet published) and libijs-0.35.so is in neither
# Hummingbird nor this factory; libgs
Previous style finding (config/factory-build-backlog.toml:354 missing pr) is fixed at 2315431.
No findings from: correctness, security, intent-alignment, style.
Confidence: 5/5 (safe)
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
@Karibhou this PR is back in your court ( Review asks:
Reply here or push a new head to this branch (same sign-off); CI and the review swarm re-run automatically and the PR returns to the normal path. Your relay session received this as a |
Hive-Run: projectbluefin#501 Hive-Plan: factory-build-backlog source-family import Hive-Spec: factory-build-backlog#closing-a-backlog-gap Assisted-by: Codex via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: Karibhou <157208370+Karibhou@users.noreply.github.com>
Assisted-by: Codex via GitHub Copilot Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Signed-off-by: Karibhou <157208370+Karibhou@users.noreply.github.com>
2315431 to
22eaef5
Compare
There was a problem hiding this comment.
Summary: Re-review of the newest revision of this change, which adds the recipe for a library that decodes compressed black-and-white scanned images. The only thing that changed since the last review is that the branch was brought up to date with the main line; the recipe itself is unchanged and the regenerated backlog snapshot still passes its consistency check. One small stale comment in a neighbouring recipe remains, as noted last time.
Delta reviewed: 23154319 → 22eaef5f. GitHub's compare shows 11 files because the branch was rebased onto main (#815); a direct content diff of the two heads shows only config/factory-build-backlog.toml and reports/factory-build-backlog.json differ, and those hunks are exactly #815's ffmpeg reconciliation from main. python3 tools/factory_build_backlog.py --check at 22eaef5f → snapshot consistent (197 entries, 189 pending); ls -d packages/*/ | wc -l = 652, matching docs/architecture.md:81,133-139.
docs-currency (low, advisory — still present from previous review)
packages/ghostscript/ghostscript.spec:61-62 says libjbig2dec.so.0 ... are in neither Hummingbird nor this factory; once this PR lands, this factory does carry it (packages/jbig2dec/jbig2dec.spec:93-95). The bcond rationale still holds via the runtime version check (ghostscript.spec:63-64,127-134), so this is not blocking — a one-line comment refresh for whoever next touches ghostscript:
# jbig2dec and ijs are built from the copies in the tarball. This factory now
# carries jbig2dec (packages/jbig2dec), but libijs-0.35.so is in neither
# Hummingbird nor this factory, and libgs
No findings from: correctness, security, intent-alignment, style.
Confidence: 5/5 (safe)
— hive: agent=reviewer backend=copilot model=claude-fable-5.1 copilot=1.0.88
|
Merged by hive contributor-lane auto-merge: CI green for ≥30m, reviewer verdict approve, @Karibhou holds the trusted tier in this hive's contributor registry. |
Summary: This prepares a library that reads compressed black-and-white images for the project's package repository. It adds verified build instructions for a missing component used by image-processing software. A successful hosted build and a test on a running system are still needed before users can rely on it.
What I need from you: Review the recipe and run the trusted hosted package build.
Changes
jbig2dec0.20 from Fedora Rawhide commit71d5c85badc25b84422e2c30ef25c157fadeaff3(treeceed5b28859b857f20a77e34a53b2f4fbd65cebb), retaining its recipe and provenance. Itslibssubpackage provides the requestedjbig2dec-libsoutput.Evidence and validation
sourcesmanifest:74d40b4d5b21dcb2b6ed15de84e9ef292774783547d0a36c7185740f3fe34341665b4aa756369d27d158404f15f36f3edd2fd1a8187a196d08c9b06ab3d04101.python3 tools/factory_build_backlog.py --check,just check, andjust testpassed; each full test run reported 1,396 passed and 2 skipped. Tests used an existing virtual environment because the system Python lacks pytest.pre-commit run --all-filesfailed on existingpackages/powerstat/packit.yamltab syntax and an existing shellcheck warning in.github/workflows/rebuild-rpms.yml.git diff --cached --checkflags three whitespace-only differences in the unchanged imported upstream spec.jbig2dec, install from Hummingbird plus the factory, and record a representative decoding result and published binary identity.— hive: backend=codex