summaryrefslogtreecommitdiff
path: root/src/host_io.zig
Commit message (Collapse)AuthorAge
* replica clients join again: one that diverged (a digest the session's unit ↵Gabriel Schneider2026-10-30
| | | | | | | | | | | | did not end on) and one the session let go to upgrade dial it again, the same join as the first, and say why in a line over their frame until a key; a client of another build is told "session upgraded to a newer pardes: upgrade this client too" tty.zig `remote` is a loop of joins over one terminal: the link ended, its reader joined, the replica let go, the session dialled again (an upgrade's exec and load, and its daemon's, retried for 20 s), its dump loaded, its view's size sent again. `Remote.notify` draws the line (`noticeLine`, the client's own, never the session's core). The welcome carries the session's semantics: the daemon judged the hello by the hash it last heard, which an upgrade under way may not have said yet. replica.zig `Peers.refuseAll`: before an Upgrade's dump each client is sent `refused` after what it was sent (up to 200 ms) and let go, its view with it. "A newer pardes is installed" is a Host call (`newer_build`) the core makes in its focus step, so replica clients take the line from the tape; the GUI still says it itself. PARDES_TEST_DIVERGE=<n> puts a client's core off the session's once quiet after its n-th unit. test/daemon.py: the clients test has one rejoin and agree; the upgrade test has a client of the old build told to upgrade and one of the build running join again and agree. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* upgrade: a newer build at the path a session started from is said at a ↵Gabriel Schneider2026-10-29
| | | | | | focus-in, once a build, the path read just after the first frame (off a startup's way); an upgraded process lets the binary it replaced go (held, its deleted file kept its blocks) and sweeps its old image's prompt files (kept for an exec that failed); test/daemon.py --upgrade asks through /ctl, checks the fds and prompt files level, and runs on FreeBSD (procstat, the daemon's registry, no kept old binary there) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* upgrade: the Upgrade word (SPC s U, a tag, 9P) moves a running session onto ↵Gabriel Schneider2026-10-29
| | | | | | the pardes installed where it started, in place, by exec at a step boundary: its dump in the DumpDir, its pty masters, terminals, stubs and daemon link kept open and named on the new command line, every view and running shell its own again; the daemon re-execs too, its registry on its command line; a build that cannot load the dump falls back to the old binary, or starts afresh with the shells adopted and the dump kept (test/daemon.py --upgrade, --upgrade --gui) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* web: a dump is a .dump by name everywhere (the download, the embedded ↵Gabriel Schneider2026-10-29
| | | | | | fixture, -l's help); the touch check resizes a pane by its grip, there being no rule between panes Co-Authored-By: Claude Opus 5.5 <[email protected]>
* FreeBSD: the tty pardes builds for it and runs there, terminals, 9P and the ↵Gabriel Schneider2026-10-29
| | | | | | | | | | | | | | | | | | | | | | | | daemon included Cross-compiled with -Dtarget=x86_64-freebsd.15.0 (15 ships libutil.so.10; Zig's default 14.0 links .so.9) and run on the FreeBSD 15.1 box: - 9p_io serves there: fstatat, not Linux's statx. madvise is Linux's. - WINCH reaches the sigwait thread through a handler that does nothing: FreeBSD drops a signal ignored by default as it is sent. - A dead terminal answers a write with ENXIO, which std does not expect: a Debug build traced it to stderr, that same terminal, and recursed until the stack ran out (a session whose ssh dropped hung). No tracing there. - waitid's FreeBSD numbers and 64-bit id: a command pane's exit is seen. - No /proc: a shell's directory and program by sysctl, ttyTaken by the foreground group, the foreground job's name by its program. - File watching by FreeBSD 15's inotify. - ghostty's kitty loader: off Linux a stat comes before a t=f open (a FIFO hung it); FreeBSD's shm_open refuses O_NONBLOCK. - vaxis's test terminal has no read off Linux: no input thread for it. - test-install: the test-build programs installed into <prefix>/test, to run where the build cannot. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* sessions register with the daemon: a terminal's session forked off its stub, ↵Gabriel Schneider2026-10-29
| | | | | | | | | | | | | | | | Detach and --attach in place for terminals and windows, pty masters kept, Adopt pardes in a terminal reaches (or starts) the daemon, opens /dev/tty and forks: the child is the session, in a session of its own (setsid), drawing on that terminal (tty.Term.fd, a raw vaxis Tty made from the fd); the parent stays as its stub, the job the shell waits on. Not the daemon spawning it from a passed fd, env and argv: the user's terminals each run in a 9ns user and mount namespace, which a session spawned by a per-user daemon would not share, and a fork costs one small process's page tables where the spawn cost an IPC and an exec. GUI windows and --detach sessions just register. tty: Detach (and a terminal's end: EOF, EIO, a hangup the stub passes on, taken on the signal thread) puts the terminal back, drops it and the stub's stdio, tells the stub (which ends, exit 0), and leaves the session running with nothing drawn (Host.watched), its messages on the clock and `size` its own. pardes --attach[=name] hands its terminal through the daemon; the session enters it as the first one, sized, queried and drawn whole, and that process is its stub until it leaves. A hangup never ends a kept session; TERM and INT do, telling the daemon so the shells go with it (one async-signal-safe send). gui: Detach closes the window (its swapchain or renderer) and keeps the device, atlas and pipelines; the loop sleeps on SDL's queue; pardes-gui --attach[=name] opens a window again. Closing a window still quits. --detach is a session started detached: the tty host with no terminal yet, the GUI with a hidden window closed at once. forkShell sends the daemon a dup of every master and the hosts tell it what they close, so a session killed outright leaves its shells running; the new Adopt word takes them in as terminal panes (the Tty word's pane, its spawn taking the pty, a 1x1 size then the pane's to make it redraw). The Attach word, which switched a session's own terminal to a detached session through the frames, goes, with its effect and request in the core. PARDES_DAEMON=off (the e2e harness's) runs a session as before. test/daemon.py: 50 detach/attach cycles at a level fd count with no stale socket, kill -9 and Adopt, and nothing left; --gui the same for a window. test/fs.py: a forked session's socket is found by its own pid, and a hangup leaves a kept session running. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* drag 2/3: a pane's or column's grip let go past the window's edge opens it ↵Gabriel Schneider2026-10-24
| | | | in a new window (Handover.zig)
* Search converge 4/6: the disk engine, Grep and Find's listing on a worker ↵Gabriel Schneider2026-10-24
| | | | (rg when there, else a sorted walk), generation-numbered batches, a debounce, a 2000-line cap and More, and the declarations parse off the loop
* One inbox for every host: the GUI's queue moves to host_io.Inbox, macOS and ↵Gabriel Schneider2026-10-23
| | | | | | | | | | | | | | | the detached server use it The inbox takes a wake callback instead of an SDL flag and owns starting a pty reader and an LSP query and delivering their answers, status and pipe results. macOS gains the GUI's rule that only a pane's current reader and the current query get in; its reap now cancels the reader on the inbox before its future, since a reader blocked on a full queue cannot be canceled. The detached server's completion mailbox goes; its socket stays the loop's wakeup. Two old macOS ABI test compile errors fixed, so its unit tests run on Linux. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Audit: the hosts' wire on std.Io Reader/Writer, Lsp and selection_pipe ↵Gabriel Schneider2026-10-23
| | | | | | | | | | deliver through a sink, std.Io clock and sleep wire.zig reads and writes through std.Io.Reader/Writer instead of its own framing helpers; host_io.Lsp.start and selection_pipe.start take a callback sink, so each shell keeps no copy of the worker plumbing. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Callstack lists every caller and pages a tree too big for one query; memory ↵Gabriel Schneider2026-10-23
| | | | | | | | | | | | | | | | bounds and resident-size fixes Callstack walks while the server answers, counts on the message row past a second, puts test callers last at each level, and ends a tree past one query's memory (memory.limits.arena.lsp_query, a 256 MiB region taken and given back per query) in a row whose Callstack shows the next page. Grep reads whole files. New hard limits: undo text per file, image decode pixels, the LSP snapshot of other open files, one paste. Safe builds no longer fill worst-case buffers with 0xAA (memory.unfilled), big regions are kept off transparent huge pages, and malloc_trim runs after big results. Also: Look's git show pages into a +Diff, and dropped marks in File. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A terminal under a wrapper is where its foreground shell is, not where the ↵Gabriel Schneider2026-10-22
| | | | | | | | | | | | | | | wrapper stays A shell that re-execs itself under a wrapper (fish's `exec 9ns --mntgen -- fish` when NINE_MOUNT is unset) left /proc/<forked pid>/cwd at the wrapper's first directory, so the pane never followed a cd and a relative path printed after one (fd's b/c.txt) looked up nothing. host_io.ttyCwd: while the forked process still runs the shell pardes started, its own directory, as before, so a job in another directory (a nested pardes, (cd sub && make)) never names the pane; once it is something else, the tty's foreground group's when that is the same shell, else no answer. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* zig fmt for Zig 0.17Gabriel Schneider2026-10-22
|
* Build with Zig 0.17Gabriel Schneider2026-10-22
| | | | | | | | | | | ghostty, libvaxis and uucode move to ghostty's 0.17 pins, cloud9 to its 0.17 main, SDL to castholm's release-3.4.x (3.4.18), zstbi and tree-sitter to forks with 0.17 build files (see build.zig.zon). A bare zig build copies the binaries into ~/.local/bin (-Dhome-install=false leaves it out): 0.17 keeps the prefix from build scripts. A terminal's edited copy goes when the program clears the screen. tinywin.snap is gone; perf baselines are re-recorded. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Add a readable static Callstack pane and align tagline textGabriel Schneider2026-10-22
|
* Edit's w to a file changed on disk since read is refused once, as Save is (w ↵Gabriel Schneider2026-10-01
| | | | | | again to overwrite), the rest of an X/'/w written and the refused named; Edit filters get a pane command's environment ($PARDES_MOUNT, $PARDES_9P, $PARDES_PID, $winid); an Edit with no commands runs beside one whose commands run unless it touches a file that one changes, refused busy (EBUSY) as is a second Edit with commands Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A look at a directory opens a pane listing it, as acme's directory window: ↵Gabriel Schneider2026-10-01
| | | | | | named dir/, its entries in acme's columns (bytewise, dotfiles kept, a directory's marked /), a look at an entry from there, Get or a look again reads it again, never dirty, Save refused; DirLook terminal keeps ls in a terminal Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Saving the init file in a pane (Config) applies its settings at once; its ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | commands are not run a second time A setting changed in the init file took effect only at the next start. The host's Save now hands a written init file's text back to the core, which applies its setting lines again (each idempotent), with the same notice and err for one that fails; a command line in it (`Newcol`) belongs to the start and is left alone. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A Rename the server answers for this file alone, while other open files of ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | | its language say the name, is previewed, never applied; Lspwhy shows the files synced first zls renaming at a declaration answers with that file's edits only, cold or warm, whatever it was told of (checked against zls directly: from a call site in main.zig it renames both files, from the declaration in util.zig it never does), so the first Rename applied in util.zig and left main.zig broken. The client now looks for the old name in the other open files of the language; when one says it and the answer left it alone, the edits are a preview with a row naming each such file, and the message row says the server renamed in this file only. Lspwhy carries the open files too, so its trace shows `synced X first` for the query it explains. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A Rename, references and the like tell the server of every open file of its ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | language first, so a rename reaches files it had not seen and the preview lists them zls renames only in the documents in its store, and syncDoc told it of a file only when a query was asked from it, so renaming in progress.zig left main.zig unchanged and broken, with no preview. For a query whose answer can reach other files (rename, references, implementation, workspace symbols, calls and type hierarchy), the host snapshots every other open file and the client syncs those its server answers for first. The mock server answers a rename in each document it was told of; lsp-rename-files opens main.rs by a look, renames from lib.rs, and its preview lists both. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* With Pager pardes, a terminal's shell also gets SYSTEMD_PAGER and ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | SYSTEMD_PAGERSECURE=0, and the pager's path is quoted only when it needs it journalctl and systemctl read SYSTEMD_PAGER, not PAGER, and in their secure mode run only less, so a pty/run of either still waited in it; and systemd, splitting a pager on blanks and exec'ing it, never ran the quoted path. Where the user's environment sets neither, a terminal now gets SYSTEMD_PAGER as PAGER is set and secure mode off (this pager has no shell to escape to), and the path is single-quoted only when it holds a blank or shell character. Command panes keep cat. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A terminal pane draws the kitty graphics its program places, so yazi's image ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | and PDF previews show yazi asks the pane XTVERSION, and ghostty-vt answers "libghostty": yazi takes it for Ghostty and sends its previews by the kitty graphics protocol with Unicode placeholders (a=T,U=1, then U+10EEEE cells). ghostty-vt stored the images, but nothing drew them: the placeholders showed as tofu in the GUI and as raw codepoints in the tty, in every build since terminal panes answered queries (vkznsvzs, 2026-07-01). Before that no query was answered and yazi fell back to chafa's text art. Each frame, TerminalImages.zig reads the active screen's images and placements (virtual ones through their placeholder runs, and ones placed on a cell), cuts each into the cell rows it covers, maps those rows through the body's own row walk, clips them to the pane, and hands them on as ImagePlaces with exact geometry (NativePlacement.exact). The GUI draws them as textures; the tty passes them to an outer terminal with kitty graphics as one placement per row, its source cropped to whole cells. Placeholder cells draw no text. Scrolling, a delete, a resize, the alternate screen and a closed pane all fall out of reading the emulator's state every frame. A pane now knows its size in pixels: CSI 14/16/18 t are answered, the emulator's width_px/height_px follow the shell's cells, and the pty's winsize carries pixels (host_io.ptyWinsize), so yazi and a nested pardes size their images right. ghostty-vt's image store is 64 MiB per screen (its library default, 10 MB, refused a large preview), PNG transmissions are decoded with stb_image, and RGB images are converted only while shown. test/yazi_preview.py runs yazi in a hidden window on a PNG and a PDF and checks the captures, also through a tty pardes nested in the pane. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* pardes - reads stdin into a +Pager pane, and is a terminal's PAGER and ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | | GIT_PAGER, Pager pardes|off to opt out; command panes page through cat A paging command in a terminal, or run through its pty/run, sat in less, and a pty/run never answered. `pardes -` now reads stdin to its end, strips terminal escapes (colour, OSC, overstrike), and shows it: inside a session as a clean <cwd>/+Pager pane made through pane/new and back at once, outside one as a new editor whose first pane it is, as `vim -` does. A terminal's shell gets PAGER and GIT_PAGER set to this pardes's own path and `-` where the user's environment sets neither; `Pager off`, a setting in /ctl, init and DumpConfig, leaves terminals started after it to the environment. A command pane's one-shot shell keeps cat, its output being a pane already. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A command run from a pane gets acme's $winid, the serial of the pane it was ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | run from; one from a column's or the workspace's tag gets none PARDES_PANE names the command pane itself, so a script a click ran had no way to address the pane it was run in, as acme's win scripts do with $winid. The command pane now remembers that pane and its child gets winid set to it; a header tag's command, which no pane owns, gets winid unset. PARDES_PANE is unchanged. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A pane's child runs with PAGER=cat and GIT_PAGER=cat where the user's ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | environment sets neither, so a pty/run or a command pane never hangs in a pager git log or man run through pty/run, or in a command pane, sat in less at its first screen: no one presses q, and the run never answers. The child's environment now names cat as the pager when the inherited one names none; a user's own PAGER or GIT_PAGER is kept. It is every pane child's, a terminal's shell too, since a pty/run's line runs in it. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* An editor's private pardes-osc133 prompt files go at Exit and at a fatal ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | | signal, and a start sweeps the ones a dead editor left /tmp had gathered thousands of private prompt files: Exit removed its own, but an editor killed by a signal, or a crashed one, left them for ever. They now carry their editor's pid in the name, a SIGTERM or SIGHUP removes them before the signal's death, and each start removes those whose pid kill(pid, 0) answers ESRCH for. Files of the older shape carry no pid and are left alone, as a running older editor may still read them; the shared ones in the runtime directory are kept, being every session's. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A background job killed while the tty probe walks it is not the tty's owner, ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | | however far into exiting it is host_io's "a background job is not the tty's owner" failed now and then, 2 runs in 20 under load, always just after `kill %1`. The probe walks the shell's children and reads each one's exe. A killed background sleep, in the moment it is exiting (its exe link gone, not yet a zombie), has no exe to read and is not a zombie to skip, so the probe called the tty taken. A process outside the foreground process group holds nothing whatever state it is in, so one whose exe cannot be read is now skipped when its group is not the foreground one. Under the same load the test passed 50 runs in a row. The unit-test step's perf gate trips under that load, which is not this test. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A detached session no client watches draws no frames: yes | head -c 50M ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | | | | | through pty/run goes from over 30 s to 11 s Profiling the 50 MB run showed 70% of its time in draw.render, recolorAnsi alone 43%, walking cells through PageList.pin. The detached session drew a frame after every pty read, with no client attached to see it. A host now says whether anyone watches (Host.VTable.watched). The detached server answers whether a client is attached, and while none is the core keeps the frame owed instead of drawing it. A /screen read draws its own, and a client attaching gets one (hello sets needs_frame). Debug, 80x24: 10 MB went from 7.36 s to 2.10 s, and 50 MB from over the 30 s client timeout (about 37 s by rate) to 11.08 s. What is left is the terminal's own parsing and scrollback growth. A session a frontend shows still draws after every read. Pacing those frames is the next step if that case matters. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The hosts' Save, Dump and DumpThemes writes and their shell reaper are ↵Gabriel Schneider2026-10-01
| | | | | | | | host_io's, once, where the tty, the gui and the detached session each had a copy writeFile, writeDump and dumpThemes were byte-for-byte the same in tty.zig, gui.zig and detached/server.zig but for how each reaches its core, and the gui's and the session's reapers the same but for a pid <= 0 against == 0 (a pid is never negative there). Each vtable entry is now a line calling host_io's. ttyTaken, fgName, killJob and ptyResize stay each shell's: they are already one call to host_io each, and what differs is how a shell reaches its pty (an optional with a file, an optional with an fd, a bounds-checked array). macos.zig keeps its copies (no macOS build here to check it). No behaviour changes. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Merge: Terminal rename + Kill signal orderGabriel Schneider2026-10-01
|\
| * Kill stops a command's shell before its jobs, so the line never runs on to ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | its next command in between signalTty signalled the running job's group first, then scanned all of /proc for the session's & jobs, then the shell. A job dead of the signal lets a job-control shell run the line's next command, and once pids wrap below the host's older processes the scan takes long enough for it to: "sleep 0.4; touch killed" killed at 150ms touched the file. The host_io test caught it every run after the pids wrapped (pid_max 4194304, new pids near 17000, 229 older processes above them to scan). The shell's group now goes first, then the session's groups, then the foreground one. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | Terminal.zig is terminal.zig: a file of functions and no fields takes a ↵Gabriel Schneider2026-10-01
|/ | | | | | | | namespace's lowercase name The last of the deferred renames, now that the tty and theme agents have landed: panes.terminal at its importers, the alias lines in Text.zig and File.zig and panes.zig's own uses following. dump.zig's Terminal struct, a dump record, is not this. The served sources list and docs/design.typ name the new file. test/perf.zig's references change, so the three perf baselines take its new harness id with their numbers as recorded. No behaviour changes. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Not-found failures say it once, in plain words: grep: pattern not found, ↵Gabriel Schneider2026-10-01
| | | | | | | | callers not found The ENOENT that 9ns reads from a failure's words came from a tacked-on suffix, grep: no match, not found, definition: nothing found, not found, no shell "x", not found. They now say it once: grep: pattern not found, find: name not found, Callers: callers not found, definition: not found, shell "x" not found (...). Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Every child starts with its signals at their defaults and none blockedGabriel Schneider2026-10-01
| | | | | | Shells, command panes, language servers, the link opener and the v9fs mount inherited the editor's signal mask (the tty's blocked SIGWINCH) and anything it ignored (a SIGHUP nohup ignored), since only handlers reset at exec. Each fork now resets every disposition and clears the mask before its exec (resetChildSignals); std's spawn for selection pipes runs with the mask cleared across it. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A path with a part over 255 bytes is refused, never a panicGabriel Schneider2026-10-01
| | | | | | std's statFile takes the kernel's ENAMETOOLONG for a bug (errnoBug), so a name, look, DumpDir or Save path with a part over 255 bytes panicked the editor (exec.kindOf via writeName, recentKeeps, dumpFailed). Every non-test statFile now goes through fs.statPath, which refuses such a name as NameTooLong first; fs.py drives long, looping, not-a-directory and not-ours paths through name, look, DumpDir, Dump and Save. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A bash or fish under another name gets its prompt marksGabriel Schneider2026-10-01
| | | | | | The shell's kind was read from its file name alone, so a renamed or differently linked bash got no marks and read busy for ever. A name that says nothing now resolves its links, then looks in the program for bash's or fish's own strings; a POSIX shell's name (sh, even when it is bash) is taken at its word. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Tty or Shell naming a file that is there but not executable says soGabriel Schneider2026-10-01
| | | | | | It said no shell "x", not found, as for a name that is nowhere. A file that exists without its execute bit now says not a shell: x is not executable. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A script whose interpreter is not there is told apart from a missing shell, ↵Gabriel Schneider2026-10-01
| | | | | | | | and Tty refuses it up front Tty's up-front check found the script and let it through, so the pane was made, its exec failed ENOENT, and the log read new, msg shell: shell not found, del, then the err. The check now reads a script's #! line and refuses it, interpreter /no/such/interp not found, and an exec that fails ENOENT on a file that is there says the same. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Errnos that fit: a Dump into a missing DumpDir and a shell that is not there ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | are ENOENT, a GUI-only setting EINVAL A Dump into a DumpDir not there failed EIO `file not found`; `no shell "x"` read as EIO through a mount; and a GUI-only setting on a terminal, EINVAL in the reply, had no word 9ns reads as EINVAL. Each says so in the words that give its errno: `no such directory`, `not found`, `invalid here`. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A shell that cannot start fails Tty and pty/ctl exec with why, before either ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | | answers A shell whose exec failed (a script's missing interpreter) was reported started: Tty and pty/ctl exec succeeded, then the pane died. The child now reports a failed chdir or exec through a close-on-exec pipe the parent reads before acknowledging the shell; the host's spawn fails with the reason (`shell not found`, ENOENT), which fails the waiting write. A Tty whose first shell never ran leaves no pane; a terminal restarted by pty/ctl exec keeps its pane. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* pty/ctl exec fails the write when the shell cannot start; a directory ↵Gabriel Schneider2026-10-01
| | | | | | | | | | | | | removed under a shell leaves its name `exec` on pty/ctl restarted a shell in a directory that was gone, the write succeeding and the pane dead; and the pane was then renamed `<dir> (deleted)` from /proc, so a restart failed even once the directory was back. `exec` is now refused up front, ENOENT, where the directory is gone; a shell the host cannot start fails the waiting write (late failure, as a Save's); and a `(deleted)` cwd is no name to take. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* pty/run's busy names the program holding the terminalGabriel Schneider2026-10-01
| | | | | | | | | busy said only that something ran. Where the host can tell (Linux, the tty's foreground group's /proc comm, a new fg_name host call), it says busy: <program> is running; elsewhere, busy as before. The GUI's host gains the one small callback (render agent's file: gui.zig, fgName). Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Tty and Shell refuse a directory: not a shellGabriel Schneider2026-10-01
| | | | | | | | | | A directory passes access(X_OK), so Tty /etc made a pane whose shell exited 127 and Shell /etc was taken. The shell lookup now refuses a directory, and both say "not a shell: /etc is a directory" (Tty and Shell share one refusal, host_io.Shell.refusal). shellset's golden takes the shared wording (re-recorded by name). Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Kill stops a command pane's whole command, its & jobs includedGabriel Schneider2026-10-01
| | | | | | | | | | With job control on, a command's & jobs get process groups of their own, so Kill's signal to the foreground group and the shell's group left them running, orphaned. A stop of a command pane now signals every process group in its shell's session (Linux; a /proc scan). Jobs still outlive a command that exits on its own. Docs and the skill say it per route. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Kill stops a command pane's whole line again, now it runs with job controlGabriel Schneider2026-10-01
| | | | | | | | | | | | With the line run under job control (the change before), the job running has a group of its own, and Kill's SIGTERM to the tty's foreground group stopped only that job: of `sleep 30; touch x` the touch still ran. For a command pane the hosts now signal the shell's group as well, so the whole line stops, as it did; a line typed at a prompt still loses only its foreground job. The test forks a real pty for both halves: a background job outliving its command and the pty's hangup, and a killed line not running on. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A command pane's background job outlives its commandGabriel Schneider2026-10-01
| | | | | | | | | | | | | | A dogfood agent's background jobs died when the command that started them finished, even under nohup. The command's shell leads its terminal's session, and on its exit the kernel hangs up the terminal's foreground group, which with job control off is the shell's and every job's. The line now runs with job control on (bash, sh, dash, zsh, ksh -m; fish status job-control full), so a job has a group of its own and lives on, printing below exit N, and survives the pane closing. While a job holds the pty the pane is not reused, so the directory's next command gets a second pane (cmdexit's golden, re-recorded by name). Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Shell prompt files are one per content under $XDG_RUNTIME_DIR, so a killed ↵Gabriel Schneider2026-10-01
| | | | | | | | pardes leaves none behind; tests remove their /tmp dirs Every pardes host wrote its own pardes-osc133-bash-* and -fish-* files to /tmp and removed them only at a clean teardown, so each killed session, test and crash left two: 72K of them had piled up. They are now written once per content, named by its hash, in the user's private runtime directory, renamed into place whole and shared by every pardes; without that directory the old private /tmp files remain. fish's -C source is quoted. The 9p_io tests remove their runtime dirs with what the listener left in them, and the snapshot runner removes its retry captures when every retry passed. Co-Authored-By: Claude Opus 5.5 <[email protected]>
*-. Trial merge: 9P + helix + renderGabriel Schneider2026-10-01
|\ \
| | * Step core animation by the shell's clock and sleep to the next wakeGabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A shell now answers Host.now, its monotonic clock in ns, and the pump advances core animation to it: one .tick per whole 16 ms frame since core time last stood still. The same animation therefore takes the same time at 60, 120 and 144 Hz, over ssh and after a slow frame (the GUI's re-armed clock ran them ~25% slow at 144 Hz; the tty's post-frame sleep drifted). nextWake() lists every core animation in one place: a frame from now while anything moves, the end of the wait while something only waits (a message lingering, the look-hover delay), null when idle. Shells sleep exactly that long, and a wait is jumped to its end in one step, so an 800 ms linger costs one frame instead of fifty. An overshoot under 1.5 ms after a step is let go: at 60 and 120 Hz every display frame takes exactly one step (a 16 ms frame against a 16.67 ms vsync would otherwise double-step every 24th). The six tick drivers are gone: tty's timer thread only times the wait, interruptibly (a newer, shorter request cuts short a sleep still timing a longer one); the GUI's AnimationClock, web's JS tick bank (pardes_tick now takes rAF's timestamp), the detached server's and the board's ticks; grid mode steps a virtual clock straight to each wake. PARDES_TEST_CLOCK, set by the snapshot harness, gives tty and the detached server the same virtual clock. Every stepped frame is drawn. The old pump never drew the last frame of a fade (a .tick asks for no frame, and the fade was over by the check), so a theme switch stopped at 9/10 of the way until the next input: theme.golden, themesel.golden and the GUI's acme-light scene move to the theme's true colours, and nothing else changes. The frozen previous grid is captured only while a panel transition is chosen, not on every idle frame.
* | | A command's exit is told once its output is in, by the pty's state, not a timerGabriel Schneider2026-10-01
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The watcher woke the host a second time 60 ms after the exit by its own clock while the grace was counted from the reap: a host busy for 10 ms missed it, and the tag said running for ever and Kill did nothing; and under load exit 0 could land before the last output. As decided, no timer: the exit is told once it is reaped and the pty says nothing is left (poll: no POLLIN, and no POLLHUP, which means the end of file is on its way behind the output), else at that end of file, checked after each chunk of output. The tty host checks inside its step so the frame shows it. The four hosts' copies are one host_io.takeExits/commandEof, which close a told command's pty at its end of file (the fd and the GUI's reader leaked when the exit came first). A finished command pane whose pty a job it left still holds is not reused, so that job is not hung up; and the reset before a reuse is SGR 0, not DECSTR, which ghostty's stream does not implement. Co-Authored-By: Claude Opus 5.5 <[email protected]>