Skip to content

Show the new Session, Window and Pane directories on perl, raku, erlang and ocaml - #331

Merged
tony merged 11 commits into
mainfrom
fix/refresh-four-ports
Nov 18, 2026
Merged

tony merged 11 commits into
mainfrom
fix/refresh-four-ports

Conversation

@tony

@tony tony commented Nov 17, 2026 •

Copy link
Copy Markdown
Contributor

libtmux.org's perl, raku, erlang and ocaml trees will show the capability directories their libraries just added to Session, Window and Pane, and perl's reference examples stop reading the test-only LIBTMUX_TEST_TMUX. These four publish from the revision of the committed API model, so their source-bound data is regenerated here at each port's current head.

Summary

Two fixes for main, which #332 left failing, so this branch can pass CI:

Test plan

  • Each code fix test-first: OCaml showcase mapping, both exporter layouts for OCaml and Raku.
  • --check for the model, api-sources, tmux-usage, manifest packages, example sources, mentions and perl POD of the four ports; check-api-links.
  • 63 focused site test files for the four ports and all of packages/api-model; the api-prose directory test covers each new directory.
  • pnpm -C site type-check (0 errors), oxlint, ruff check and ruff format --check.
  • test (test-all) on this branch, including the type-link ceiling for these ports.

tony added 8 commits November 17, 2026 01:58
…their operations

why: libtmux-ocaml master adds three CI-executed showcase programs to
examples/showcase.json, options, panes and clients. The extractor attaches each
whole program to the API symbols it demonstrates through a fixed table and
stops with "Invalid OCaml whole program" on a capability the table does not
name, so the model could not be generated from master at all.

what:
- Attach the options program to `Client.set_option`, `Client.set_environment`
  and `Client.set_hook`, the panes program to `Client.swap_panes`,
  `Client.move_pane`, `Client.break_pane` and `Client.respawn_pane`, and the
  clients program to `Client.switch_client_at`,
  `Client.resize_control_client_at` and `Client.request_detach_client_at`, all
  in `Libtmux_unix`
- Test that each program reaches its operations without its region markers,
  and that a capability outside the table is still refused
…when the ref has it

why: libtmux-ocaml moved its Python export wrapper from tools/export_api.py to
scripts/export_api/export_api.py in 3eb3122, and the shared port-docs workflow
still called the old path, so a build of any ref from that commit onward failed
at the export step before the model was read. Refs from before the move
(tags and older branches) still keep the wrapper under tools/, and the same
workflow builds those too.

what:
- Pick scripts/export_api/export_api.py when the selected source has it and
  fall back to tools/export_api.py otherwise; the dune build of
  tools/api_export.exe is unchanged because the exporter program stayed in
  tools/
- Run the producer test against both layouts, including a failing wrapper, and
  require the wrapper to be the path that exists in the source

Source refs: libtmux/libtmux-ocaml@3eb3122
…n it from there

why: libtmux-raku moved its RakuDoc exporter from tools/export-docs.raku to
scripts/export-docs/export-docs.raku in f6b61be. The shared port-docs workflow
still ran the old path, and the Raku adapter accepted only that path as the
exporter named in the artifact's provenance, so neither a build nor a model of
any ref from that commit onward could succeed. Refs from before the move keep
the old path and the same workflow builds those too.

what:
- Run scripts/export-docs/export-docs.raku when the selected source has it and
  tools/export-docs.raku otherwise, after the dependency install
- Accept either of the two paths as the exporter in the artifact's provenance
  and refuse any other
- Test the workflow step against both layouts with stand-in zef and raku
  commands, and the adapter against both paths and a third

Source refs: libtmux/libtmux-raku@f6b61be
…directories

why: The Perl reference pages for LibTmux::Session, LibTmux::Window and
LibTmux::Pane cannot show the "What is on here" capability directory until the
model carries the prose it is built from, and the reference examples still ran
through the author fixture that reads the test-only LIBTMUX_TEST_TMUX. The
committed model still described 388f27171. libtmux-perl#24 adds the three
directories and libtmux-perl#26 rewrote the programs the examples are taken
from as self-contained ones that start their own server on a private socket;
both are on master.

what:
- Pin Perl to libtmux/libtmux-perl@1f8332193 and regenerate its API model and
  path data (exported with Perl 5.44.0 and PPI 1.291, the toolchain recorded
  in the previous model; the exporter digest is unchanged), tmux usage, source
  cache, guide cache, example sources and manifest packages with the
  repository's generators; the mention index and the sidebar sidecar come out
  unchanged
- Adopt the source change between the pins: the model keeps 634 symbols with
  none added or removed and 8 docs changed. LibTmux::Session, LibTmux::Window
  and LibTmux::Pane gain the directory, and the examples of LibTmux::Server,
  LibTmux::Selection, LibTmux::Workspace::load, LibTmux::Async::Server and its
  snapshot method become the self-contained programs, so none of the 634
  symbols carries LIBTMUX_TEST_TMUX any more
- Keep every reviewed tmux usage claim: all of them hold at the new revision,
  so only the usage file's revision binding moves
- Take the three changed guides into the guide cache: the filtering and
  compatibility guides and the examples README
- Move the cached examples/workspace.pl to the new revision
- Expect the directories of Session (3 rows, 6 references), Window (3 rows,
  8) and Pane (4 rows, 13) in the directory test, and that no Perl reference
  example reads LIBTMUX_TEST_TMUX
- Move the POD-link tests from LibTmux::Pane to LibTmux::Client: Pane.pm now
  differs from the file the 0.003 release shipped, so its POD link names the
  release and says the page documents other source, which the tests assumed
  of no other module; Client.pm is still the release's own file
- Leave the landing example, its recording and the home fixture at
  8a9ccf32c0a5: examples/ and scripts/lib are unchanged between 8a9ccf32c and
  the new pin

Source refs: libtmux/libtmux-perl master
1f8332193ef4e51523860339ecac5b614bb679ad (#24), previous pin
388f27171ee6256dc7dbaef6f878d394f9ca474a; self-contained examples in
libtmux/libtmux-perl@8a9ccf3.
…pability directories

why: The Raku reference pages for the session, window and pane handles cannot
show the "What is on here" capability directory until the model carries the
prose it is built from. libtmux-raku#19 adds the directory to
LibTmux::Model::SessionRef, WindowRef and PaneRef, and the committed model,
guide cache and manifest still described ae97f9f87. The pin is the merge commit
of the PR, which is on master.

what:
- Pin Raku to libtmux/libtmux-raku@cce868d81 and regenerate its model and path
  data (exported with Rakudo 2026.09, the compiler recorded in the previous
  model; the exporter now lives at scripts/export-docs/export-docs.raku and its
  digest moves), tmux usage, source cache, guide cache and manifest packages
  with the repository's generators; the mention index, the example sources and
  the sidebar sidecar come out unchanged
- Adopt the source change between the pins: the model keeps 1069 symbols with
  none added or removed and 3 docs changed, the three handle classes that gain
  the directory
- Re-anchor the reviewed tmux usage rules: the directories added 230 lines
  above the shared transfer and constructor code of Model.rakumod, so the two
  citations of the same-pane check in `PaneRef.join` and `PaneRef.move` and the
  five captured-view constructor reviews now name the line the text is on;
  no claim or rule changed
- Take the rewritten examples README into the guide cache
- Expect the directories of SessionRef (6 rows, 16 references), WindowRef
  (6, 18) and PaneRef (11, 38) in the directory test; the captured views
  Session, Window and Pane carry no directory in the source
- Leave the landing example, its recording and the home fixture at
  9a9838047f15: examples/quickstart.raku is unchanged between that revision
  and the new pin

Source refs: libtmux/libtmux-raku master
cce868d81823db54a0d161a94db4d078d09f1861 (#19), previous pin
ae97f9f87b73db63d4ab3cabb108a195f64e4f6f.
…ux_pane capability directories

why: The Erlang reference pages for tmux_session, tmux_window and tmux_pane
cannot show the "What is on here" capability directory until the model carries
the prose it is built from. libtmux-erlang#24 adds the directory to the three
module docs, and the committed model, guide cache and manifest still described
33fd7fe98, the head of the tmux_server directory PR. The pin is the merge
commit of the PR, which is on master.

what:
- Pin Erlang to libtmux/libtmux-erlang@48aacd01a and regenerate its model and
  path data (exported with OTP 29.1.2 and the same exporter digest as before),
  tmux usage, source cache, guide cache, manifest packages and mention index
  with the repository's generators; the example sources and the sidebar
  sidecar come out unchanged
- Adopt the source change between the pins: 325 symbols become 329
  (`tmux_pane:wait_exit/2` and the types `exit/0`, `wait_exit_options/0` and
  `wait_exit_result/0`) and 3 docs change, tmux_session, tmux_window and
  tmux_pane, which gain the directory
- Name the line of the five dispatcher citations of
  `classify(tmux_server:command(Server, Command, Options))`: the new
  `tmux_target:submit/3` repeats that line below `run/4`, so the text alone no
  longer picks one; the citations stay on `run/4`, line 90, and no claim or
  rule changed
- Take the six changed guides into the guide cache (README, installation,
  jobs, compatibility, options reference and API), and the one new mention of
  `tmux_pane:wait_exit/2` in the operations guide into the mention index
- Link the 18 programs of the examples index at the new revision; the
  repository's examples/ gained no program, `terminal_job.py` became the
  escript `terminal_job.erl`, which the index never listed
- Expect the directories of tmux_session (3 rows, 12 references), tmux_window
  (3, 12) and tmux_pane (7, 35) in the directory test
- Leave the landing example and its recording at 539c7297ceae:
  examples/readme_demo.erl is unchanged between that revision and the new
  pin, and the displayed rebar.config is a consumer's own, not the
  repository's

Source refs: libtmux/libtmux-erlang master
48aacd01add3256181ac57cfb29ae3047b93a39e (#24), previous pin
33fd7fe985454e375d5ac8533043cabf4a0b2fa5.
…ns capability directories

why: The OCaml reference pages for Libtmux.Session, Libtmux.Window,
Libtmux.Pane and Libtmux.Operations cannot show the "What is on here"
capability directory until the model carries the prose it is built from, and
the Libtmux_unix.Client directory showed five rows because one of its lists
ended its table early. libtmux-ocaml#18 adds the four directories and words
that list as references alone, and the committed model, guide cache and
manifest still described ffab5fe, the last revision the generators read before
master added three showcases and moved the exporter. Both are handled now, so
the pin moves to master.

what:
- Pin OCaml to libtmux/libtmux-ocaml@1611f141a and regenerate its model and
  path data (exported with OCaml 5.5.2 and odoc-parser 3.2.2, the toolchain
  recorded in the previous model; the exporter now lives at
  scripts/export_api/export_api.py and its digest moves), tmux usage, source
  cache, guide cache, example sources and manifest packages with the
  repository's generators; the mention index and the sidebar sidecar come
  out unchanged
- Adopt the source change between the pins: the model keeps 1141 symbols with
  none added or removed. The four modules gain the directory, the Client
  directory returns to eight rows, and ten operations carry the new options,
  panes and clients programs besides the fifteen that already carried a
  program, whose source link now names the new revision
- Keep every reviewed tmux usage claim: all of them hold at the new revision,
  so only the usage file's revision binding moves
- Take the three changed guides into the guide cache: the showcases, testing
  and benchmarking guides
- List the three new programs, showcase_options.ml, showcase_panes.ml and
  showcase_clients.ml, in the examples index, link all 26 at the new revision
  and expect 26 in the index test
- Name the workflow step "Execute the standalone showcase programs" in the
  seven capability panels: ocaml.yml split the step that ran the installed
  documentation and the showcases in two, and the panel's command
  `opam exec -- python3 test/showcase_test.py "$TMUX_BINARY"` is the second
- Expect the directories of Session (5 rows, 13 references), Window (6, 15),
  Pane (7, 21) and Operations (7, 49), and eight rows, not five, in that of
  Libtmux_unix.Client
- Leave the landing example, its recording and the home fixture at
  ffab5fecd298: examples/showcase_graph.ml is unchanged between that revision
  and the new pin, and the dune files the home page shows are its own

Source refs: libtmux/libtmux-ocaml master
1611f141a1f69bbfd396b15ff12715ff260543dd (#18), previous pin
ffab5fecd298d9516f181576231f54926630118b.
why: The exporter-path fallback compared the artifact's producer.path,
typed unknown, against the accepted paths; packages/api-model's
type-check (TS2345) failed #331's test.

what:
- Require producer.path to be a string before EXPORTER_PATHS.includes
tony added a commit that referenced this pull request Nov 17, 2026
why: The exporter-path fallback compared the artifact's producer.path,
typed unknown, against the accepted paths; packages/api-model's
type-check (TS2345) failed #331's test.

what:
- Require producer.path to be a string before EXPORTER_PATHS.includes
@tony
tony force-pushed the fix/refresh-four-ports branch from 737d611 to 0360d28 Compare November 18, 2026 00:35
… as happy-dom's

why: The sheet-mode test from #332 cast its matchMedia stub to
the DOM library's MediaQueryList, which astro check rejects against
happy-dom's Window.matchMedia; the error failed main's test job at
f772074 and every branch based on it.
what:
- Cast the stub to ReturnType<typeof win.matchMedia>

(cherry picked from commit 740e1e0)
tony added a commit that referenced this pull request Nov 18, 2026
…its drawn height

why: #332 gave every picker trigger a touch-target min-height below the
phone breakpoint. In the header that row wraps under the bar at 35rem,
so the bar grew past the 97px the responsive-header check allows at
390px, and the docs test job failed on #331.
what:
- The compact and coarse min-height skips the header's trigger
- Under pointer: coarse the header's trigger takes the same hit-area
  pseudo-element as the header's chips, so its tap is 44px tall while
  the bar keeps its height
…its drawn height

why: #332 gave every picker trigger a touch-target min-height below the
phone breakpoint. In the header that row wraps under the bar at 35rem,
so the bar grew past the 97px the responsive-header check allows at
390px, and the docs test job failed on #331.
what:
- The compact and coarse min-height skips the header's trigger
- Under pointer: coarse the header's trigger takes the same hit-area
  pseudo-element as the header's chips, so its tap is 44px tall while
  the bar keeps its height
why: Since #332, pickers open as a bottom sheet on phones in portrait and
a side sheet on short landscape screens (design record 0071). Two
browser checks still assumed a popover: the picker-filters check wanted
7px from every edge at 320 x 360, where the sheet meets the edges by
design, and the header-menu check closed a picker by clicking its
trigger, which the sheet covers.
what:
- check-picker-filters: bounds reads the panel mode and keeps the 7px
  inset for popovers only
- check-header-menu-sections: a sheet closes with Escape; a popover still
  closes from its trigger

This branch was previously deployed

2 inactive deployments
docs-preview-cleanup — 1670e7d6 Deployed Nov 18, 2026 by tony via cleanup #258
docs-preview — 1670e7d6 Deployed Nov 18, 2026 by tony via publish-preview / publish #1485
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant