Copying ::punk::libunknown::epoch verbatim into a fresh interp/thread
carries per-epoch 'added' maps that drive zipfs_tclPkgUnknown's
can_skip_sourcing short-circuit - sound only in the interp that performed
the scans. Packages the donor obtained at an earlier epoch (or with no
unknown call: Thread is baked into thread::create'd threads) are absent
from the donor's current-epoch map, so recipients skip their pkgIndex
sourcing and requires fail where stock tclPkgUnknown would succeed.
Bit deterministically as repl_console_backend_tkconsole under the suite
9.0.5 runtime: 'package require Tk' bumps the epoch pre-init, the
codethread inherits the post-bump sparse map, the code interp then can't
find vfs/Thread/registry, its init aborts before the G-001 conredirect
junction vars exist, and the first pending-collect error left the repl
reader inactive until the driver failsafe. Tcl903 escaped only by an
accidental extra epoch bump in its code interp (fresh map, full rescan).
Fix: punk::lib::libunknown_epoch_export (PUNKARGS-documented) - epoch
record with every 'added' map emptied; counters/untracked/scripts/index
globs preserved. Used at all four propagation sites (interp_sync,
snapshot, repl %lib_epoch%, repl safe-arm). Hardening: the repl
pending-collect now tolerates missing junction vars so a code-interp init
failure can't wedge the reader.
Also: test socketpairs pinned to -myaddr 127.0.0.1 (all-interfaces
listens raised windows firewall consent popups for each newly built test
exe - user datapoint).
Verification: consolebackends 3/3 both runtimes (suite 42s->14s); full
runtests sweeps identical on suite runtime and Tcl903: 1037/1021/15/1
with the sole failure exec-14.3 (documented machine baseline). G-098
acceptance sweep met. Pre-existing PUNK_TEST_TK opt-in quirk noted in the
goal (silent early end, both runtimes, out of scope).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
New src/vfs/punk9wintk905_beta.vfs modelled on punk9wintk903.vfs: suite-built
tk9.0/tklib0.9/tcllib2.0/tcllibc/thread3.0.1/vfs1.4.2 payload (tk + thread
dlls in bin/ per the ../../bin pkgIndex convention; tclvfs dll rides inside
vfs1.4.2 - vfs.tcl loads from $::vfs::self), plus twapi5.0.2 + sqlite3.51.0
carried over from the 903 vfs for basic human trial. tcludp deliberately
omitted (one vendorlib_vfs.toml target line away if -teelog trials wanted).
Runtime = _build/suite_tcl90/out/bin/tclsh90szip.exe copied (untracked) to
bin/runtime/win32-x86_64/tclsh905_beta.exe - static tclsh 9.0.5 with the
tcl library in its attached zip; the zip wrap merge_overs that into the kit
so tcl_library survives (same mechanism as tclsh902z->punk902z).
Built with 'make.tcl vfs -confirm 0': punk905_beta.exe INSTALLED; kit smoke
passes - Thread/vfs/vfs::zip/sqlite3/twapi/tcllibc(md5 C-accelerated)/tk
9.0.2 all load with functional pokes, punk boot machinery active.
Beta-kit policy recorded in mapvfs.config + goal: wrap fresh suite runtimes
as *_beta kits even before punk tests pass, for early interactive trial.
Future noted: 'make.tcl vfs <kitname>' single-kit build + configured-kit
listing; move tclvfs/thread into the runtime-attached zip.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
suite.tcl 'test' is now a gate: full tcltest output logged, totals parsed
(all.tcl's exit code does not reflect failures), failed test names compared
against expected_test_failures.txt - unexplained failures fail the action;
stale baseline entries are noted on full runs. Totals on core-9-0-branch
9.0.5 / zig 0.16.0: 67281 run / 55835 passed / 11435 skipped / 11 failed,
GATE PASS, deterministic across two full runs. The 11 are dispositioned in
the baseline: 6 fCmd (OS username vs profile-dir machine trait), 2 filename
(admin-share self-globbing), 2 safe-stock (install-layout accessPath
tokens), 1 console-bug-f10d91c2 (powershell driver yields empty result in
batch context; semantics verified independently by G-076 kill procedure).
Runtime package plumbing gaps closed (built artifacts without install
wiring): thread pkgIndex generated (lib/thread3.0.1 -> bin dll), registry
dll install artifact hooked to install-libraries (was orphaned - registry
1.3.7 + dde 1.4.6 now load), tclvfs script package installed with vfs.tcl
+ pkgIndex.tcl generated from .in templates (vfs 1.4.2 + vfs::zip 1.0.4).
punkshell runtests under the suite runtime: checkbugs 5/5, opunk
backends.test PASS (Tk-widget console suite runs unskipped on the
suite-built Tk), repl consolebackends 2/3 - repl_console_backend_tkconsole
open (code-interp console_pending_out junction init missing; DRIVER-DONE
timeout-failsafe).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
tcllib pure side via its own installer (explicit -pkg-path; driver now unsets
TCLLIBPATH/TCL_LIBRARY/TK_LIBRARY for hermetic child shells after the default
libdir followed inherited TCLLIBPATH to C:/TCLPKGS - additive pollution
verified and removed). install-libraries now installs the whole tcl library
tree (its own pkgIndex references msgcat/opt/... subdirs the piecemeal
install missed). critcl pinned 3.3.1 (git) + tracked critcl_zig.config
(zig cc -target x86_64-windows-gnu); driver replicates sak's direct critcl
invocation over the 21 declared files -> tcllibc.dll at out/lib/tcllibc
(provide 2.0). Evidence: md5 accel(critcl)=1; sha1 5MB critcl 12.3ms vs
pure 3.06s (~248x) - the punkcheck hashing pain quantified and fixed.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Installer behaviour checked first per user caveat: module exclusions,
app-script generation, doc options, and the aggregating top-level
pkgIndex.tcl (gen_main_index) - not a straight copy; critcl DSL stubbed to
no-ops (pure-tcl). Driver runs it with the suite-built tclsh in batch mode,
packages only, landing out/lib/tklib0.9. Smoke: package require Tk +
tooltip -> 2.0.4 in the suite shell.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Per user direction and draft refinements 2026-07-20: matching tklib alongside
Tk (pure-tcl, own installer.tcl where it does more than straight copy);
tcllib installed via its installer machinery (installer.tcl/sak.tcl checked
for non-copy behaviour) with critcl-built tcllibc driving zig cc; the
totals-parsing test gate owns the 72-failure triage. goals_lint clean.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
The build016 experiment port lands in the tracked recipe (root_module/
createModule creation, addLibrary+linkage, .root_module.* method sweep,
linkSystemLibrary options arg, array_list.Managed, Io.Dir dir-walks with
b.graph.io, wrapfiletofile ported to the 0.16 entry/Io idiom). 0.14 support
deprecated per user direction; 0.14.1 remains archived on punkbin. Verified
on the pinned default (bin/tools/zig-x86_64-windows-0.16.0): build PASS,
shells 9.0.5, tcl-test parity 557/357/200/0 identical to the 0.14.1 build.
punk-getzig payload pins -> 0.16.0 (windows zip + linux tarball), re-wrapped
checkfile ERROR-free, ps1 twin refreshed. punkbin commit 54714d7 holds the
0.16.0 archive; push pending gitea recovery (server currently 500s).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
vqtcl was an old-tclkit experiment with no confirmed public upstream - wiring,
helper module, staging and docs removed (future tclkit work: TEMP_REFERENCE/
metakit + KitCreator). New 'test' action runs tests/all.tcl on the suite-built
static shell with -testargs passthrough; verified assocd/util/lindex: 557 run,
357 passed, 200 constraint-skipped (no C test commands in a release shell),
0 failed. install-libraries tm filenames synced to the tree's provided
versions (tcltest-2.5.11, http-2.10.2, platform-1.1.1) - the 2024-era names
made 'package require tcltest' fail on a filename/provide version conflict.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Recipe: b.host -> b.graph.host; redundant second UNICODE tclMain add into the
objs lib removed; exe's ansi tclMain compiled with -UUNICODE -U_UNICODE
-DTCL_ASCII_MAIN (zig cc windows-default defines made it a duplicate UNICODE
compile under 0.14.1 - lld duplicate Tcl_MainExW/commons). Driver: per-item
recipe staging, per-zig-version local caches (.zig-cache-<ver>; caches must
not be shared across zig versions), zig resolution -zig > PUNK_ZIG > pinned
bin/tools/zig-x86_64-windows-0.14.1 > PATH. Verified: PASS on 0.14.1 AND
regression PASS on 0.14.0-dev.2074; pinned-default no-option run PASS.
Pinned archive minisign-verified and published to punkbin (abcd87c).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Scriptset (src/scriptapps/bin) and the generated bin wrapper renamed; the
untracked bin/getzig.ps1 twin replaced by punk-getzig.ps1 (byte-copy of the
.cmd - launchability convenience for cmd.exe-or-powershell starts). Payload
improvements: self-references updated, and the ziglang.org release URL now
derives its version from the single $releasearchive pin (was a separately
hardcoded 0.15.1 vs the 0.15.2 archive). Re-wrapped via scriptwrap multishell;
checkfile ERROR-free (the two boundary-scan warnings are the documented-normal
polyglot ones). bin/AGENTS.md gains the punk- naming policy section
(exceptions: dtplite/sdx/kettle wrappers, getpunk; sweep of bits/runtime/
tclargs/selfsign is G-097).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
First real buildsuite: build905.zig (verbatim G-076-validated recipe) +
helper modules + tools, driven by suite.tcl which stages the recipe into
src/buildsuites/_build/suite_tcl90, arranges sources (tcl fossil
core-9-0-branch --nested with FOSSIL_HOME isolation; thread/tclvfs trunk;
zlib from the tcl tree's compat/zlib; vqtcl from a local dir), generates the
configure-products fresh checkouts lack (tclUuid.h, tclsh.exe.manifest,
thread.rc include), builds and smoke-tests. Verified PASS: static + zip
shells report 9.0.5 (checkout 1a9c3b9d96), Thread 3.0.1 builds from current
trunk (new generated-include fix in build_tclthread) and loads. The 2024
TCLSH_PIPEREPL/tclZipfs/tclCmdIL patches are now tracked under
suite_tcl90/patches with a semantics README (not applied by default builds).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Blanket /src/buildsuites ignores removed from .gitignore and
.fossil-settings/ignore-glob; retired /vendorbuilds/ ignores removed with
them. src/buildsuites/_build remains ignored via the existing _build globs.
samplesuite1 sketch marked as such, buildprefix moved to the _build
convention (vendorbuild concept retired per user decision 2026-07-20).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Sweep: G-096 detail and the punk::lib gate comment now point at the archived
path with achieved phrasing. goals_lint clean (72 active, 25 archived).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
punk::repl (pending 0.5.2): repl_handler's full 'chan configure' probe is now
guarded - on runtimes carrying the upstream f10d91c2d3 fix a dead console is
DELIVERED as a readable event and the probe errors ('couldn't read console
mode: broken pipe'); previously this backgrounded and left the repl idle with
its reader inactive. Terminal input loss now routes through new
repl::console_input_lost - the same teardown console_watchdog uses (helper
shared by both paths; watchdog probe and message unchanged).
punk::lib (pending 0.5.0): check::tclbug_console_deadspin_fixed_in = 9.0.5
after the G-039 kill procedure passed on a zig-built core-9-0-branch runtime
with the watchdog disarmed (clean script-visible eof exit 0.2s after conhost
kill). One edit flips both consumers: 'help tcl' warning and watchdog arming.
Regression-verified on affected 9.0.3: bug=1, watchdog armed, rescue at 1.9s
via the shared teardown. Suites: console+repl+lib families 115/115.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
The stale copy (2026-07-19 20:55 build) survived three make.tcl bootsupport
runs: two in-session runs failed the copy with 'invalid argument' (destination
held open - a running shell's zipfs mount on the bootsupport modpod, the
G-087-era self-lock class) and the user's post-shell-close rerun left it
untouched (cause unobserved - if that run reported no copy attempt rather
than a failure, the punkcheck staleness records having absorbed the earlier
failed attempts would be worth noting for G-094). Direct file copy from the
current root modules build (22:13, md5 b1226b3d) succeeded once the lock was
gone; bootsupport now matches root modules and _vfscommon.vfs (already fresh
via the earlier vfscommonupdate).
Claude-Session: https://claude.ai/code/session_01BNUVVkYq9vHa6G3S9a3XTZ
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Goal/Acceptance wording updated per user approval (verification runtime need
not be an official release). Upstream f10d91c2d3 fix confirmed merged and
present in TEMP_REFERENCE/tcl9; both G-039 defects addressed. Progress
entries: toolchain policy (zig-only for project integration), known-good 2024
zig build recipe determined and duplicated (zig-out90dup2026), 9.0.5
verification runtime built (tcl905 checkout, build905.zig: fresh-checkout
preconditions tclUuid.h + tclsh.exe.manifest, mingw_unicode_entry_point for
upstream's removed TCL_BROKEN_MAINARGS). Remaining: G-039 kill procedure on
the 9.0.5 runtime.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Content-identical rename (flip edits landed in the previous commit); GOALS.md
entry removed and the summary record (amended *** acceptance) appended to
GOALS-archive.md per the achieved flip workflow. Rides along: CHANGELOG.md
gains the required [0.17.1] entry matching the punkproject.toml bump from the
previous commit (punk::path 0.4.0 *** pathglob + separable lattice; runtests
targeting rework).
Claude-Session: https://claude.ai/code/session_01BNUVVkYq9vHa6G3S9a3XTZ
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
In-place closeout edits ahead of the archive move (per the goals/AGENTS.md
pure-rename discipline): detail-file Status flipped to achieved 2026-07-20
(all amended acceptance clauses verified - see the detail file's Progress
record), the console-attachment finding pushed to live G-061's Notes as part
of the archive reference sweep, and punkproject.toml patch-bumped 0.17.0 ->
0.17.1 for the punk::path 0.4.0 pathglob addition shipping in kit payloads
(user-directed). goals_lint transiently red until the archive-move commit.
Claude-Session: https://claude.ai/code/session_01BNUVVkYq9vHa6G3S9a3XTZ
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
make.tcl modules + bootsupport + vfscommonupdate after the punk::path ***
source changes: path-0.3.0.tm -> path-0.4.0.tm snapshots in
src/bootsupport/modules and src/vfs/_vfscommon.vfs. The runtests parent
resolves punk::path from bootsupport - the refreshed snapshot is what enables
*** targeting at the runner discovery layer (discovery.tcl requires 0.4.0-).
Known ride-along not included: the bootsupport templates-0.2.0.tm modpod copy
failed to refresh ('invalid argument' - destination held open, likely a
running shell's zipfs mount; the stale copy predates this session) - reported
separately.
Claude-Session: https://claude.ai/code/session_01BNUVVkYq9vHa6G3S9a3XTZ
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
User-approved contract amendment replacing the day-1 file-relpath filter: one
directory-oriented pathglob semantics at every layer, with the ergonomics gap
closed in the syntax itself.
punk::path 0.4.0:
- *** as a whole segment matches ZERO or more segments (join-aware regex in
pathglob_as_re: an adjacent separator is absorbed into an optional group),
so X/*** = X and everything below - the subtree spelling; ***/f.txt matches
at any depth including the root.
- include separability fix: an exact or single-segment-glob full match no
longer sets allbelow (the deep walk previously dragged the whole subtree in
for X and X/* patterns; the zipfs walk was already separable).
- exclude separability fix (new exclude_state classifier, both walks): only
**|***-tail patterns prune the matched subtree; other matches exclude just
that directory's files - the **/_aside vs **/_aside/** argdoc distinction is
now real. Dead _path_matches_any removed.
- legacy bare-* include collapse removed (a * in -include-paths is the
one-below lattice form, not a match-everything alias; no caller used it).
- globmatchpath default-nocase driveletter exception rewritten as a direct
regex-head rewrite (the regex-text file-split broke on c:/***'s glued group
and accepted non-alpha drives).
- path.test +5 lattice/*** pins (40 total).
runtests.tcl:
- -include-paths/-exclude-paths pass through to treefilenames unchanged;
argdocs teach the separable X | X/* | X/** | X/*** lattice; directory and
file-name axes independent (tails); single-file targeting = exact dir
pattern + file tail.
- -serial-paths reconciled to the same directory-of-file reading (default now
modules/punk/console/*** modules/opunk/console/*** via
runtests_discovery::dir_matches_any) - the pre-existing file-path serial
matching was the other same-option-different-referent divergence.
- per-pattern zero-match stderr advisory (the X/**-for-subtree foot-gun points
at X/***).
- json report emitter: whitespace-padded entier-looking values now take the
quoted-escaped path ('string is entier -strict' accepts surrounding
whitespace - a failed test's "0\n" result_was embedded a raw newline and
truncated the parseable report line).
- testsupport/discovery.tcl now a thin passthrough (requires punk::path
0.4.0-); pathdiscovery.test rewritten to the directory-semantics contract
(22 tests incl. the X/** zero-match advisory characterization); AGENTS.md
targeting guidance swept to /*** across src/tests and src/modules/opunk.
Verification (native tclsh 9.0.3, piped, serialized): 45-probe matrix, path
suite 40/40, runner suite 22/22; vs the pre-G-093 baseline the default
discovery FILE SET differs by exactly the added runner suite and result
parity only by the five new path tests (+ that suite); mode parity -jobs 16
vs singleproc: PARITY ok (94/1037/failed=1 exec-14.3). Goal contract amended
in the detail file; full record there.
Claude-Session: https://claude.ai/code/session_01BNUVVkYq9vHa6G3S9a3XTZ
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Root cause (recorded in the goal file): punk::path::treefilenames matches
-include-paths against the DIRECTORY containing files, and its documented
X/** boundary semantics exclude files directly in X - so suites keeping
.test files at the subtree root (scriptwrap, binscripts, punkexe) were
silently missed by X/** targeting: the G-092 four-pattern (2/13 files) and
repeated-flag (zero files) failures. punk::args -multiple accumulation
probed healthy - repeated flags and a single space-separated value reach
discovery identically; no punk change was needed or made.
runtests.tcl now filters the tailbase-relative FILE paths after an
unrestricted walk (new testsupport/discovery.tcl: globmatchpath,
any-include-match, exclusion wins): X/** includes files directly in X, an
exact relative path targets a single file, a bare directory path matches
nothing (append /**). -exclude-paths exposed (same syntax and accumulation,
exclusion wins over inclusion). -discover-only prints the discovered list
plus a RUNTESTS_FILES tcl-list line and exits before any run machinery -
subsecond via a two-stage boot (discovery phase loads only punk::args +
punk::path; punk/Thread/shellrun/punk::tcltestrun and the testinterp/child
path assembly load after the exit).
New regression suite runner/testsuites/discovery/pathdiscovery.test
(23 tests: pure filter, fixture-tree walk+filter, end-to-end -discover-only
children including both G-092 characterization forms; ~7s wall,
parallel-safe). src/tests/AGENTS.md targeting wording reconciled; new
src/tests/runner/AGENTS.md. G-093 index status proposed -> active
(user-directed).
Verification (native tclsh 9.0.3, piped Bash context): pre-change baseline
default -jobs 16 (93 files / 1010 tests / failed=1 exec-14.3 only) vs
post-change -jobs 16 with -exclude-paths runner/**: PARITY ok; post-change
default run differs from baseline only by the added suite (+23 passing);
mode parity -jobs 16 vs singleproc sequential: PARITY ok (94/1033/1).
Claude-Session: https://claude.ai/code/session_01BNUVVkYq9vHa6G3S9a3XTZ
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
make.tcl modules + bootsupport + vfscommonupdate after the stage 4/5 source
changes: commandset layout 0.1.1->0.2.0 and project 0.3.0->0.4.0 snapshots in
src/bootsupport and _vfscommon.vfs; thin-layout sync refreshed the store
layouts' make.tcl copies and the templates modpod payload (repacked
templates-0.2.0.tm) for the retired marker exclusion comment.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Content-identical rename (edits landed in the previous commit); GOALS.md entry
removed and the summary record appended to GOALS-archive.md per the achieved
flip workflow.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Stage 4 - documented overlay semantics implemented (commandset::layout 0.2.0):
layout_materialize resolves a layout to its overlay chain (vendor tier then
custom tier per store+vendor, <name>@<base> derived-layout folders expanding
their base chain first, cycle-guarded) and composes multi-folder chains into a
disposable staging dir applying <item>.anti deletion markers (zero-byte marker
files recommended - committable, unlike the originally documented empty marker
folders, which are also accepted). Single-folder chains with no markers are
used in place, so punk.project (store and zipfs modpod) behaviour is unchanged.
project.new (commandset::project 0.4.0) copies from the materialized path and
the layout listing/scanning commands show the effective merged view.
Fixtures: punk.othersample is now a live derived layout (new custom-deck
fauxlink ref + .gitignore.anti demo marker); extra.txt/store README rewritten
to the implemented semantics. New tcltest suite
src/tests/modules/punk/mix/testsuites/layout/materialize.test (20 tests).
Anti markers deliberately stay a plain filename convention rather than
tag-bearing fauxlinks: a fauxlink is a reference tooling may follow (fcat
auto-follows), a marker is a tombstone with no target.
Stage 5 - retirements and generated-project version seeding:
- The empty unimplemented child-side src/PROJECT_LAYOUTS_CREATE_FROM_... marker
folder is deleted along with its project.new and make.tcl sync exclusions.
- The vendor/custom x _project matrix collapses for layouts: custom/<vendor> is
the one override level; _project folders join no overlay chain (the
pseudo-vendor remains for decktemplates module/utility templates only).
- project.new seeds generated projects post-materialization (only-if-missing,
never from payload): punkproject.toml ([project] name + 0.1.0), a matching
CHANGELOG.md ## [0.1.0] header, and punkorigin.toml - an interim generation
provenance baseline (utc, layout name + resolved store target,
punk::mix::templates version, punkshell version when derivable, origin path)
for the G-027 pull design to adopt (noted in G-027).
Also: script-mode/programmatic project.new robustness (explicit shellrun
require; punk::mix::util askuser/do_in_path/foreach-file fully qualified -
the util:: shorthand exists only after cli commandset import). Includes the
concurrently-landed (other session, previously uncommitted) 0.3.1 early
module-name validation fix in project.new, shipped within the 0.4.0 bump.
G-087 flipped achieved 2026-07-19 (verification in the detail file; archive
move follows in the next commit). punkshell 0.16.0 -> 0.17.0.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Outcomes of the 2026-07-19 punkcheck module review:
- G-094 [proposed]: single OO record lifecycle - installtrack/installevent
becomes the sole implementation with an eager|deferred persistence policy,
punkcheck::install rebuilt as an OO consumer in deferred mode, the
zero-caller installfile_begin/finished/skipped trio retired to shims, and
the o_record_list staleness fixed. The single save chokepoint is G-095's
attachment point.
- G-095 [proposed]: concurrent-writer safety - atomic temp+rename saves,
advisory EXCL lockfile serializing whole installer events per punkcheck
root, merge-on-flush for deferred mode, non-interactive-safe
duplicate-INSTALLER recovery. SQLite recorded as rejected for now (TDL
human readability/debugability overriding; three-part reconsideration bar).
- G-027: acceptance reframed - .punkcheck records demoted from sole
provenance basis to acceleration layer; safe-overwrite classification
survives record loss via a committed baseline record (origin identity +
pulled version, pristine content re-derivable from the origin's VCS),
degrading conservatively (report/confirm - never silent overwrite, never a
wedged pull). Approach classification bullet extended to match.
- G-095 notes carry the tiered expendability taxonomy (build trees: recopies
only; synced-targets bin/ deploy: non-self-healing skip wedge; derived
projects: G-027 baseline) and the cross-tier invariant: absent records
degrade to skip/confirm, never to overwrite.
- (The G-087 cross-pointer to G-094/G-095 already landed with the stage-3
sources commit.)
goals_lint clean (73 active-index goals, 22 archived).
Claude-Session: https://claude.ai/code/session_01L7MfuYBfYpUFpVafo5JowC
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
vendor/punk/project-0.1 loses its stored src/bootsupport snapshot (10.6MB, 254
files) and src/vfs/project.vfs lib+modules snapshots (7.7MB, 230 files);
vendor/punk/basic loses its 2023-era bootsupport snapshot. Thin layouts keep
structure, scripts, docs, fossil/git settings and now carry the three
bootsupport modules*/include_modules.config manifests plus README markers
documenting generation-time injection (bootsupport README, project.vfs lib and
modules READMEs).
custom/_project deleted entirely (punk.project-0.1, punk.shell-0.1,
punk.basic) together with their two decktemplates _overrides refs: these were
bootsupport-snapshot + make/build-script overlay fragments kept current by the
retired make.tcl sync - generation-time injection provides exactly that
content, so they were fully redundant (punk.shell-0.1_overrides was never a
working standalone layout).
The store payload store drops from ~43MB to under 1MB (bin/sdx.kit 300KB
remains, grandfathered - excluded from the module-shipped payload copy).
Claude-Session: https://claude.ai/code/session_01WS3VakZo4sgbFzcQFW3twZ
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
punk::cap::handlers::templates 0.3.0: module-pathtype layout refs resolve
against the providing module's own templates/project_layouts store (sibling of
its layout_refs folder) instead of the never-existing <modulebase>/src tree -
module-carried thin layouts become usable.
punk::mix::templates 0.2.0: ships the thin punk.project layout payload inside
the modpod (payload synced from src/project_layouts by make.tcl - see below);
shellproject_multivendor declaration fixed decktemplates -> src/decktemplates
(shell-project layouts/templates were invisible when cwd was outside any
project).
punk::mix::commandset::project 0.3.0: project.new injects bootsupport into the
generated project at generation time via new lib::bootsupport_inject - manifest
(include_modules.config) driven module resolution from the generating shell's
project bootsupport tree, or a standalone kit's tcl:™️:list paths (999999.*
dev versions never injected; latest-version non-glob semantics matching
make.tcl bootsupport; adjacent non-.tm support files such as struct sets_*.tcl
and textutil .tex data carried along; bootsupport lib* trees copied from a
shell project, notice emitted for manifest-less standalone kits); punkcheck-
recorded with plain-copy degradation. -layout now prefix-matches the live
layout set instead of the package-load-time snapshot.
punkcheck 0.4.0: start_installer_event no longer writes literal '$eventid'/
'$rel_source'/'$rel_target'/'$config' into event headers (braced-literal
expansion bug, latent since inception - collided as duplicate ids when install
and installtrack writers shared one db); sources with no common root vs the
punkcheck folder (e.g //zipfs:/ module-mounted layouts) are recorded as
independent absolute paths instead of erroring.
punk::mix::base 0.1.1 -> 0.1.2 (manually versioned rename): cksum_path no
longer hangs forever on vfs-mounted files - tcllib -file digest modes read via
fileevent+vwait and zipfs channels never deliver fileevents (empty
ZipChannelWatchChannel in core); non-native files are slurped and digested in
data mode (new cksum_data_command per algorithm + cksum_crc_data/
cksum_adler32_data helpers; exec-based sha3 errors cleanly for vfs paths).
make.tcl: custom/_project layout bootsupport-sync retired (project layouts
store no snapshots); sync_layouts retargeted to thin vendor layouts (make.tcl,
build.tcl + bootsupport manifest copies; punkcheck db at
src/project_layouts/.punkcheck) plus store->modpod punk.project payload sync
(excludes bin/sdx.kit and src/PROJECT_LAYOUTS_* marker; fossil settings kept
via core-exclude override); bootsupport copy onto a modpod the running
make.tcl has mounted records OK when target content is already identical
(FAILED-record churn fix); workflow text updated (DIAGRAM 1b) and help/argdoc
texts adjusted.
bootsupport include_modules.config: added previously-recordless-but-required
entries cmdline (fileutil dep), struct::set (make.tcl accelerators), textutil,
textutil::expander, textutil::patch - generated projects' make.tcl now boots
under a plain tclsh.
Docs: root/src/bootsupport/punk AGENTS.md updated for the retired sync and
build-managed modpod payload; goal detail G-087 stage-3 progress record;
CHANGELOG 0.16.0 + punkproject.toml bump.
Claude-Session: https://claude.ai/code/session_01WS3VakZo4sgbFzcQFW3twZ
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
New scriptlib/developer/nslist.tcl backing the new cross-framework tcl-nslist
agent skill (.agents/skills/ canonical, .claude/skills/ synced copy): the
discovery counterpart of whatis.tcl - lists what a namespace/package actually
exposes at runtime via punk::ns::get_ns_dicts (the n// machinery): child
namespaces with loadable-package markers, commands classified by kind,
punk::args-documented commands, -synopsis whole-namespace usage lines,
-pathcommands namespace-path resolution. Auto-package-requires unloaded
packages and follows ensemble commands to their implementation namespace.
The tcl-whatis skill/script cross-reference it (whatis redirects with a hint
when the queried name is a namespace, not a command).
punk::ns 0.9.1: get_ns_dicts packagetails/packageprefixes initialization moved
inside the per-namespace loop - with multiple matched namespaces (path globs,
trailing **) every dict after the first carried the union of all earlier
namespaces' entries, mis-marking non-package children in nslist/n// underline
styling and in the returned data; a child with several deeper packages below
it was appended to packageprefixes once per matching package (now once).
Pinned in nslist.test (nslist_classification_multins_package_buckets_reset).
Documentation-only PUNKARGS added for get_ns_dicts (returned dict keys
documented) and nslist_dict; parsing stays manual.
Root AGENTS.md: introspection note + scriptlib/developer inventory updated.
Project version 0.15.1 + CHANGELOG entry. punk::ns suite 102/102 pass.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
The multiproc child bootstrap no longer preloads shellrun + the no-op
runx warmup: 17 test files gained their missing explicit package
require lines (punk::ansi for the bare ansi suites, punk::args for the
bare args suites incl dev/dynamic-cache, punk::lib for the lib suites
and for grepstr/untabify/examples whose punk::ansi/punk::args commands
reach punk::lib internally, plus textblock for examples.test - caught
by the empirical lean run: punk::args only soft-requires textblock and
punk::args::helpers::example calls textblock::join). Children now load
only tcltest plus each file's declared deps; the singleproc testinterp
keeps shellrun as the runner's own capture mechanism.
Module-web finding recorded in src/tests/AGENTS.md and the test-file
comments rather than fixed: punk::ansi and punk::args call punk::lib
without requiring it, and a hard back-require is blocked by punk::lib's
own hard require of punk::args (cycle) - deps declared at test level;
a considered module-level resolution remains open.
UDP watch-mode note folded into child_test_runner.tcl per user
direction: no interaction today (the parent relays child output at file
granularity; verified post-lean - events + relay flow); any future live
per-child tee must be conditional on PUNK_TEST_UDPTEE so unwatched
children stay lean.
Verified (native tclsh 9.0.3): full-suite PARITY ok for lean -jobs 16
vs the pre-lean baseline, between repeated lean -jobs 16 runs, and
lean -jobs 1 vs -jobs 16 (989/973/15/1, exec-14.3 sole failure).
Timings: -jobs 16 70.7s -> 49-53s; sequential multiproc ~5m30 -> 4m32;
singleproc unchanged. AGENTS scaling guidance refreshed.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Three tests fail only under a half-attached console (stderr=console,
stdout=file): grepstr highlight tests emit empty SGR and nslist marking
loses its markers - punk::console process-global colour state under
mixed tty-ness. Fully-piped and -jobs runs green; fully-interactive
unverified. Recorded on G-061 (user direction) as a concrete
context-matrix workload for the pseudoconsole harness, alongside the
interim Start-Process selective-redirection repro technique and a
pointer to the related capture_stderr utf-16le fix (913ea749).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
User request after console-context failures went unnoticed: compact
per-file status words, failure/warning list entries, the per-file
had-failures/passed/child-failed notices, the markdown
failures/warnings file lists, the tally Passed/Failed segments and a
coloured overall STATUS on the DONE line are now SGR-coloured (green
pass, bold red fail/error, bold yellow warning).
Gating: -colour auto|on|off (default auto) colours only when stdout is
a real windows console, detected via twapi GetConsoleMode on the
STD_OUTPUT handle - the isatty-equivalent; get_console_handle proved
unsuitable (it succeeds for piped children of console shells, verified
both ways). NO_COLOR honoured in auto. Machine output (RUNTESTS_RESULT,
json) is never coloured; piped output stays plain for agents (verified
zero escape bytes piped, codes present with -colour on). Raw literal
SGR is used rather than a+/a so the runner's indicators are independent
of punk::console's process-global colour state (the same 2026-07-19
console runs showed that state shifting under mixed tty-ness). The
previously UNCONDITIONAL raw-red errorInfo block in the file-error
branch now goes through the same gate, so piped error output is cleaner
than before.
Also sharpens the -jobs guidance wording: subtree runs (e.g 28-file
modules/punk/args/**: ~50s sequential vs ~11s at -jobs 16) benefit from
-jobs - only single-file/few-file runs do not.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
dynamic_bad_tag_warns_once failed (count 0) ONLY in console-attached
singleproc runs (user report; reproduced deterministically with stderr
on a real console and stdout redirected): when stderr is backed by a
windows console channel, a pushed chan transform receives the console
driver's utf-16le wide-char byte stream - every ascii char arrives
NUL-interleaved, so content regexps never match. The punk::args warning
itself emits correctly, exactly once, inside the capture window (proven
with marker writes: markers and warning all captured, all wide).
Piped/file stderr and -jobs child processes deliver plain bytes, which
is why every piped run passed.
Fix in all three capture_stderr helpers (punk/args dynamic.test,
punk/ansi grepstr.test, punk/ns corp.test - identical latent defect):
if the captured data is NUL-interleaved, decode it as utf-16le
(fallback 'unicode' for tcl 8.6) before returning. Verified: the
console-attached args subtree run is green (246/243/3/0); piped runs
unchanged.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
User-measured reference-machine numbers 2026-07-19: -jobs 8 ~79-86s,
-jobs 16 ~61s (~5.6x vs ~5m40s sequential), plateau by -jobs 24 (~60s).
Guidance now recommends -jobs 16 and records the plateau anatomy
(slowest single child ~31s runtimecmd pair, serial console tail ~16s
piped / larger under a real console, ~10s runner overhead).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Title broadened in both tiers; Goal/Acceptance gain the -exclude-paths
clauses: expose treefilenames' existing capability on runtests (or an
equivalent post-discovery globmatchpath filter), accumulate semantics
matching -include-paths, exclusion winning over inclusion, a
subtree-exclusion run discovering exactly default-discovery minus that
subtree, and both options pinned by the same regression-suite
combination matrix. Notes record the bundling and the middle-ground
motivation examples.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
The '### runtests DONE at <ts>' human line gains '(elapsed <N>s)', the
RUNTESTS_RESULT summary line gains a trailing elapsed_seconds=<N>
key=value, and the json report gains a top-level "elapsed_seconds"
field - matching the name already used by the watch-mode run-end event,
which now shares the same computed value. Additive only; the parity
tool reads named keys and is unaffected (verified on all three report
styles).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
User-directed from the G-092 measurement findings: a single
-include-paths value with four deep patterns matched only one subtree,
and repeated -include-paths flags matched zero files, both under the
default *.test tail glob - contradicting the documented accumulate
semantics. Context records the full repro matrix (multi-pattern forms
with explicit tails worked; standalone globmatchpath matched the same
patterns, implicating the treefilenames walk/prune or option-value
shape), the candidate root causes, and the tail-glob workaround. Notes
record that treefilenames already implements -exclude-paths unexposed
by runtests - the planned middle-ground option should be pinned by the
same regression suite.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
multishell.test (10 tests, ~105s solo child - the -jobs floor) split five
ways along shared-fixture boundaries, every test moved verbatim with its
provenance lines: multishell.test keeps wrap + structure/LF-only + exec
smoke (incl WSL); multishell_wrapverify.test (fresh-wrap checkfile) and
multishell_wrapdeterminism.test (byte-identical re-wrap) each re-pay the
wrap fixture; runtimecmd_checkfile.test and runtimecmd_roundtrip.test
guard the committed bin/runtime.cmd and its source round-trip with no
wrap fixture at all. argparsingtest.test (17 tests, ~32s) split into
timeit/discover/contract/first_call (11) + argparsingtest_compare.test
(compare_*, 6). Assessed and left unsplit: dtplite.test (24.6s solo),
scriptexec.test (12s); exec.test excluded as imported core material.
File headers point at the src/tests/modules/AGENTS.md punk/mix bullet
for the sibling map; the runtests longest-first weight table carries
measured post-split weights.
runtests_parity.tcl gains a -names mode: multiset comparison of
'- testname:' lines from '-report markdown -show-passes 1' captures,
file attribution ignored - verbatim moves compare identical, lost/
duplicated/renamed tests are reported.
Verified (native tclsh 9.0.3): repeated -jobs 8 full-suite runs 1m26.3s
/ 1m26.8s (was 2m16s at G-091, ~5m40s sequential - 4.07x vs the
pre-parallel baseline, 36.6% vs the G-091 reference); slowest child
30.7s / 31.3s (bound 45s); NAMES identical (989 occurrences) vs the
pre-split sequential capture; tallies identical in every run (989/973/
15/1, exec-14.3 sole failure, files 87 -> 92); PARITY ok -jobs 8 vs
-jobs 1 and between repeats; repo clean after runs. Notable finding
recorded in the goal file: child-wall contention inflation (~45% with
50s+ dominant children present) collapsed to ~15% once the work
distribution was flattened - splitting reduced both the critical path
and the contention itself.
G-092 flipped to achieved 2026-07-19 and archived (GOALS-archive.md
record, detail file with full verification evidence to goals/archive/).
goals_lint clean. Also noted in the goal file, out of scope: runtests
-include-paths multi-pattern/repeated-flag matching misbehaved during
measurement (single value with four deep patterns matched one; repeated
flags matched zero) - contradicts documented semantics, needs its own
investigation.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
User-approved draft from the G-091 verification data: multishell.test's
~2m child wall is the -jobs floor (full-suite -jobs 8 = 2m16s); splitting
the long suites (multishell, dtplite, argparsingtest, scriptexec if
worthwhile; exec.test excluded as imported core material) along
shared-fixture boundaries targets no child wall above 45s and at least a
25% full-suite improvement, with name-level test conservation, provenance
lines travelling with moved tests, weight-table updates, and parity
gating. Context records the fixture-duplication tradeoff (total CPU up,
wall down) and the complementary lean-children candidate.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
- User Preferences: never open/update fossil checkouts with tip (newest
check-in on ANY branch); name the branch explicitly and verify the
fossil status tags line. Prompted by TEMP_REFERENCE/tcl9 being parked
on manifest-in-rc-only since 2026-07-08 via a tip checkout race.
- Child DOX Index: add missing TEMP_REFERENCE/ entry (VCS-ignored,
machine-local reference area; tcl9 nested fossil checkout stays on
trunk). The local TEMP_REFERENCE/AGENTS.md carries the refresh
procedure but is unversioned by design.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
-jobs N (implies -singleproc 0; explicit -singleproc 1 conflicts) runs
test files through a tpool of N concurrent child processes with
per-child capture files and tcltest -tmpdir subdirs, longest-first
static-weight submission, and a -serial-paths serial tail (default: the
console-state suites modules/punk/console/** + modules/opunk/console/**;
piped-child-driven suites proved parallel-safe and each serial file
costs its full child boot strictly after the parallel phase). Cached
results feed the unchanged processing loop in discovery order, so
reports are identical in shape and order to sequential runs. Watch mode
emits file-collected events (completion order, output relayed there);
file-end (tallies) stays discovery-order. The runner now reports jobs
phase times and the ten slowest child WALL times (child wall includes
boot + file-level fixture cost that per-test usec misses).
Two Tcl-core-level behaviours found and worked around in the child
spawn (see the verification record in the archived goal file):
- concurrent BLOCKING execs from worker threads convoy on Windows (all
return at the longest concurrent child's exit; plain sleepers 2-12s
from 8 workers all walled at 12.0s) -> spawn via 'exec ... &' and
poll per-pid tcl::process status; blocking fallback retained for
runtimes without tcl::process (tcl 8.6).
- Tcl_ReapDetachedProcs runs on every exec and purges finished
background children process-wide, racing the poll (nondeterministic
'key <pid> not known in dictionary') -> tcl::process autopurge false
+ explicit per-pid purge after the status read.
Verified on native tclsh 9.0.3: repeated -jobs 8 full-suite runs
2m16.2s / 2m15.6s with PARITY ok vs -jobs 1, between repeats, and vs
the pre-parallel baseline (987-test identity, exec-14.3 sole expected
failure); 2.58x vs the ~350s sequential baseline (2.41x vs one
cooler-machine 5m28.8s sequential run - recorded in the goal file);
repo clean after every run.
G-091 flipped to achieved 2026-07-19 and archived (entry to
GOALS-archive.md, detail file to goals/archive/ with the verification
record; reference sweep updated runtests.tcl comments and the
src/tests/AGENTS.md marker). goals_lint clean.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Convert punk::config from a manually-versioned module (config-0.1.tm) to
the standard magic-version scheme (config-999999.0a1.0.tm +
config-buildversion.txt, real version 0.2.0) per user direction -
punk::config was never in src/modules/AGENTS.md's manual-versioning
exception list (punk::libunknown, punk::mix::base only). No callers pin
the old version; built/bootsupport/vfs copies remain 0.1 until the next
make run, as usual.
Includes the user-authored registration of PUNK_TEST_UDPTEE in the known
punk env vars table (runtests UDP watch mode / runx -teelog, a65f9e5b -
see src/tests/AGENTS.md). CHANGELOG 0.15.0 entry extended accordingly.
Verified: package require punk::config resolves 999999.0a1.0 under the
dev module paths (package prefer latest); focused runtests spot run
green. The env vars table is materialized inside the config init proc,
so registration is exercised at shell boot rather than package load.
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
Both user-approved 2026-07-18 from the runx -teelog / runtests watch-mode
work (a65f9e5b) and the -singleproc 0 completion (218fbb1c).
G-090 records the probe evidence deliberately not addressed in that work:
float-placement starvation of a tee under the float-locked capture
junction, insert_transform re-instantiation churn, parked-worker adoption
never re-applying settings (silent cross-target log misrouting), the
per-file teehandle worker-thread leak in unwatched singleproc runs, and
the tcludp Tcl 9 receive-fileevent first-datagram-only behaviour.
G-091 is the phase-2 parallel test plan over the phase-1 child spawn:
tpool -jobs N, per-child tmpdirs, serial group for console-sensitive
suites, parity + >=2.5x acceptance (floor ~90s, multishell.test bound).
Assisted-by: harness=claude; primary-model=claude-fable-5; api-location=anthropic.com
