Repository navigation
docs(debt): correct stale, contradictory and misplaced entries - #14085
LeandroLP23 wants to merge 7 commits into
Conversation
MarcZX10
left a comment
MarcZX10
left a comment
There was a problem hiding this comment.
Reviewed — changes requested
This cleanup and open PR #11575 both remove the same libSceHttp2 default trust debt entry, but the checklist says Depends on: none and no landing order is recorded.
What I ran
Reviewed head 2fe975165876fb5f295720a24976b4ba5001c6cc against PR base ca073ec3ad0d766ea9f4658e418a5c977001a353; the merge preview against that current main was clean. Environment: Linux x86-64, Python 3.11.2, GCC 12.2.0. This documentation change was not built.
python3 tools/pr_precheck.py --repo boykopovar/AnyPS5 --pr 14085 --output-dir /tmp/anyps5-14085-review
gh api repos/boykopovar/AnyPS5/pulls/11575 --jq '{state:.state,head:.head.sha,base:.base.sha}'
gh api repos/boykopovar/AnyPS5/pulls/11575/files --paginate --jq '.[] | select(.filename=="docs/dev/debt/libSceHttp2.md") | .patch'
gh api repos/boykopovar/AnyPS5/pulls/14085/files --paginate --jq '.[] | select(.filename=="docs/dev/debt/libSceHttp2.md") | .patch'Results
| Check | Result |
|---|---|
| Diff check | Passed. |
| Conventions and PR template | Passed. |
Merge preview against current main |
Clean. |
| Live overlap check | Confirmed both open PRs delete the same debt bullet; #11575 is open at 6ae1f9381bc8feb9f91cd1af41bc07d8895ace81. |
| Build/tests | Not applicable to the documentation edits; no build was run. |
| ABI and test-quality helpers | Could not run because those workspace-only scripts are absent from the isolated PR snapshot. |
| Hosted CI | Conventions, eligibility, intake and labels passed. Windows was still running; portability/report were in progress. Product and relinker jobs were skipped. |
Unedited local check output
All conventions checks passed.
The diff check exited 0 with no output. The precheck mergeability report says the PR preview against current main was clean.
Findings
Blocking
The PR deletes the libSceHttp2 default trust bullet in docs/dev/debt/libSceHttp2.md; open #11575 deletes that same bullet in its current diff. The contribution rules say authors should agree which overlapping change lands first and declare that order with Depends on. Please coordinate with #11575's author, then either remove the duplicate deletion from one PR or record the agreed dependency/order in the checklist. This finding concerns only the duplicated documentation edit; I found no need to change the other cleanup items for this reason.
Non-blocking
None.
Sources
- #14085 diff from current base
ca073ec3to head2fe97516 - #11575 diff from base
d2863131to head6ae1f938 - Branching and overlap requirements at the reviewed base
Not verified
I did not validate every behavior statement in the edited technical-debt entries against its implementation. The review finding is limited to the exact duplicate edit and repository overlap rule. I did not run a build or product tests.
To get this to a good state
Record an agreed landing order with the corresponding Depends on entry, or leave the shared debt bullet to the other PR. I will recheck the overlap and rerun the same no-build checks on the updated head.
|
Thanks. I left that bullet to #11575: I rechecked the other overlaps AI-assisted: yes (Claude Code) |
Re-checked — overlap resolved, full review pendingThe current head no longer changes |
|
Other open pull requests touch the same code. Merging one will break the other, or both implement the same thing. If this one builds on them, list them in Depends on.
Some comparisons could not be completed because a pull request conflicts with the base branch. Review these shared files manually before concluding that there is no overlap.
|
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…rkload entry Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…1575 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
9709235 to
9e3a21c
Compare
|
@LeandroLP23 #14247 overlaps this PR on one line: the DS src2 operations entry in #14247 makes those three opcodes translate instead of throwing, so the "throw" in your rewording would be wrong once it lands, and that sentence becomes:
Would you leave that bullet to #14247, as you did with the AI-assisted: yes (Claude Code). |
What
A pass over
docs/dev/TechnicalDebt.mdanddocs/dev/debt/*.mdagainst the current code. It corrects entries that are stale, contradict another entry, or state facts the code no longer matches. Each change below was checked against the source it describes. No behaviour changes; the only code edit removes one// unknown datacomment, which the no-comments rule does not allow.Stale (the code changed since the entry was written)
libSceSystemService.md: removed "libkernel'ssceKernelConvertUtcToLocaltime,sceKernelConvertLocaltimeToUtcandsceKernelGettimezonestill use UTC". They use the host zone now (libkernel/Time/Time.cpp,HostZoneAt).libSceCesCs.md: removed "the KS X 1003 profile is not implemented".sceCesRefersUcsProfileKsX1003and itsTilde0x7eform exist, and the next entry describes them.libSceAgcDriver.md: the "Color export type against the target format" entry is removed.DecodeColorFormatnow gives number format 4 theUintpacking, and exports other than 1/2/3/9 into a UINT target throw, as the "UINT color targets" entry says.libSceAgcDriver.md, image storage tables: atomics no longer fault through a table. Only packed and BY2/BY4 tables are unsupported (ResourceTracker.cpp), asshader.mdalready says.libSceAgcDriver.md, shader float mode: dropped the claim thatFLOAT_MODEfollows a single mode. f32 min/max/min3/max3/med3 read the denormal field.shader.md, f32 denormals:DenormPreservefor f32 where the device offers it (SpirvModuleSetup.cpp), so the entry no longer says it requests none;ds_add_f32follows the field;shader.md, LDS float add: the same "does not request" clause is fixed.shader.md, DS src2: the stated reason ("not modelled for LDS float atomics yet") is obsolete, sincedsAtomicsets both flush flags. The f32 src2 forms still throw, so the reason is now that their mode behaviour is not measured.relinker.md:libkernel_web.prxno longer "exports only two functions". It has 43 throwing inventory stubs.libkernel.md:__prognameis no longer an "unknown data export". It points to the constanteboot.bin, whichgetprognameand_getprognamereturn.Inaccurate
libSceAudioOut.md,sceAudioOut2SetSystemDebugState: the entry now matchesAudioOut2System.cpp.0x80268001, a null param0x8026800C, and a call before initialization0x80268006.0x20with size 4, and a null inner param, throw.libSceRegMgr.md: the store is never filled, since the setters throw, so every valid query misses.libSceSsl.md: only libSceHttp2 callsSslTrustSystemRoots_nid_no_patch.libSceBluetoothHid.md:GetDeviceName/GetDeviceInfothrow; the ioctl forwarding describes the native module.libSceFiber.md: the host-stack window applies only to contexts below 64 KiB. Fibers are never registered with ulobjmgr, whatever flag0x1says.libSceAgcDriver.md, quad lists: only indirect quad-list draws throw. The restart bit is ignored and no recipe is recorded.libSceAgcDriver.md, unused pixel programs: DecodeState applies the alpha-to-mask override to the dithered word but not to the undithered one.libSceAgcDriver.md, PS4 display buffers: the tiled path is not reached yet, because VideoOut refuses non-linear PS4 buffers before they get there.libSceAgc.md, ACB workload calls: they no longer claim to write "the same packets the DCB calls write".sceAgcDcbSetWorkloadStreamInactivethrows, as the next entry says.libSceAgc.md: thesceAgcCbMemsetExclusiveentry no longer says "the signature is unknown". It is declared with the canonical signature, as the later entry says.libSceHmd2.md: theFBUEhH9tQ1Alink points toInventoryExports.cpp, where the symbol is.libSceAvPlayer.md: thesceAvPlayerSetTrickSpeedlink points toPlayer.cpp, where the zero-speed ordering is.libs.md: the struct isAvPlayerVideoEx, notSceAvPlayerVideoEx.Duplicates, empty entries, misplaced entries
libSceNpCommon.md: removed a second StreamReader/StreamWriter entry that a merge brought back. It contradicted the current one onSkipData.libSceAgc.md: the one-linesceAgcAcbAtomicGds_0900NID entry is folded into its full entry.libc.md: the three 128-bit division entries are folded into the compiler-rt entry that already covers them, with a link toRuntimeSupport.cpp.libSceVoiceChat.md: the "Unknown function info" entry no longer repeats the 25 handler and event exports listed under "Silent stubs".libSceCommonDialog.md(with its otherwise empty section) andlibSceLibcInternal.md.TechnicalDebt.md:sceNetCtlGetIfStatentry moves tolibSceNetCtl.md.libkernel_sys.md:shader.md:image_loaddecodes them;v_log_f32entry.Not changed here, for follow-ups:
fchmodatwithAT_SYMLINK_NOFOLLOWon a symbolic link is a declared silent stub, whilelchmodthrows for the same case.sceAgcDcbSetWorkloadStreamInactivecould write the ACB form's 9-dword NOP.pop_frontplaceholder, whose plain symbol name still resolves to the stub onmain.Overlap
tools/pr_overlap.pylists 37 open pull requests that edit some of the same lines, mostly by rewriting a whole entry for their own change, so this conflicts textually with whichever lands first. I left out three entries that open PRs are rewriting for their own work: thelibSceLibreSsl3exclusion list, which #10452, #10490, #10491, #10493, #12315, #12754, #12945 and #13897 each extend (it names 7 of the 13 implemented exports today);sceSystemServiceLaunchWebBrowser, which #12895 implements; and the Cp1254/Cp1258 "as in Cp1250" wording, which #13898 changes. The stalelibSceHttp2"default trust" entry, which #11145 made obsolete, is left to #11575, which deletes it.Tested
tools/check_conventions.py --base origin/main: 0 errors. Itsdebt-lengthwarnings are all on entries that were over 1000 characters before this change; none of them grew past the limit here.Checklist
main; no other open PR implements the same functionsNotImplemented_nid_no_patch); silent stubs are listed under Silent stubs in their library's technical debt file.mdfiles or images added to the repository (attach them to this PR)