summaryrefslogtreecommitdiff
path: root/src/ninep/ctl.zig
Commit message (Collapse)AuthorAge
* /index shows names as the log does, one line of UTF-8, and a newline in a ↵Gabriel Schneider4 days
| | | | | | | | | | | | name as \n /index printed a name as it was, so a directory with a newline in it split its row in two, and bytes not UTF-8 went through; the log turned the newline into a space, silently another name. Both now show a name one way: a newline `\n`, other controls, DEL and C1 a space, bytes not UTF-8 `\xNN`. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Every EINVAL a write gets says why; DEL in an exec or look line is a control ↵Gabriel Schneider4 days
| | | | | | | | | | | | character An empty name, a word written to log other than follow, a data, sel, dot or dirty write to a pane with no text, a truncate other than to zero, and a control character in a look line written through a held open were refused with no words, logged as `Invalid argument`. Each says its reason now; and DEL is refused in exec and look lines as in names. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* pty/ctl exec fails the write when the shell cannot start; a directory ↵Gabriel Schneider4 days
| | | | | | | | | | | | | 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]>
* A pane's word written to /tagexec or a column's exec is refused, not done at ↵Gabriel Schneider4 days
| | | | | | | | | | | the focused pane Undo, Msg and the other pane words written to a tag no pane owns ran at whichever pane had the keyboard. They are refused as the root's ctl refuses them, EINVAL; the tag's own words (a column's New, Tty, Find, Grep) still run. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A command from a missing directory fails ENOENT with no pane; one whose ↵Gabriel Schneider4 days
| | | | | | | | | | | | | shell never started ends exit 127 Exec made a command pane in a directory that was not there, whose shell never started: it said running for ever, a follower waited for an exit that never came, and Kill found it. The line now fails up front, as Tty does, `exec: <dir>: no such directory`; and a command pane whose shell the host could not start ends `exit 127` at once. A shell that could not start in its directory says `directory missing`, not `file not found`. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A shell given a directory is not taken to be in this process's before it has ↵Gabriel Schneider4 days
| | | | | | | | | | | | | spoken, so it logs no rename to the session's and back Each host read /proc/<pid>/cwd right after forking the shell and at every frame after; before the shell's chdir that says this process's directory, a rename to it, then the real one a frame later, a rename back. A shell given a directory is named by it already (newShell) and is no longer asked at fork; and until it has printed anything, an answer naming this process's directory is not taken. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A terminal is typed into only when idle at an empty promptGabriel Schneider4 days
| | | | | | | | | A look at a directory, or an exec at a terminal, typed its line after whatever sat unsent at the prompt and ran the two joined (`echo TYPEDls`). A terminal with text at its prompt now takes no line; a look at its directory opens a new terminal, an exec runs as a command pane. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Grep and Find count the rows they found, so a hit opens its +Search and a ↵Gabriel Schneider4 days
| | | | | | | | | | | | miss fails Round 18 made a search that found nothing fail, but grep and find never counted their rows (only a pane search did), so every Grep read as a miss. Both count the lines they wrote now; a Find or Grep that finds nothing fails alike, `find: no match, not found` / `grep: ...`, ENOENT, and opens no empty +Search. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A new terminal is named by its directory from the start, so its shell ↵Gabriel Schneider4 days
| | | | | | | | | | arriving there is no rename Tty's terminal was `new N /` until its shell reported its directory, then `rename N <dir>`: a rename for no change of place. The pane owns the directory it is started in from the first; only a move elsewhere renames. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A builtin failure over something not there answers ENOENT: Unmount of a name ↵Gabriel Schneider4 days
| | | | | | | | | | not mounted `Unmount nothere` failed EIO with `not mounted`. It says `no such mount nothere` now, and a builtin's failure that says `no such` (this, Tty's `no such directory`) answers ENOENT, the errno 9ns reads from those words. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Grep with no hit says grep: no match, and a pane directory that is not there ↵Gabriel Schneider4 days
| | | | | | | | | | | is passed over Grep walks every pane's directory; one not there (a scratch named into a directory not made yet) failed the whole search with `file not found`. Such a root is passed over now, and a search that finds nothing says `grep: no match` and opens no empty +Search. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A builtin that needs its argument, written bare to an exec, is EINVAL as ↵Gabriel Schneider4 days
| | | | | | | | | | through ctl Through a ctl the bare word was refused up front (EINVAL); through exec it ran, failed with EIO or asked for the argument. The exec now refuses it the same way before it runs. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* get on a pane of a peer out of reach says dial failed, as look and Save doGabriel Schneider4 days
| | | | | | | | get's failure named the path and `cannot be read` for any error but a missing file; a peer that did not answer now reads `get: <path>: dial failed: no answer`. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A rename alone does not make a pane dirtyGabriel Schneider4 days
| | | | | | | | | Any rename marked the pane dirty (saved revision one back), while the guards ask only whether its text was ever edited: /index said dirty for a pane Exit let go without asking. The name is no edit now, so /index, the guards and Restore agree; Save writes under the new name all the same. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* The root exec's refusals before a line runs name the pane, and say invalid, ↵Gabriel Schneider4 days
| | | | | | | | | | | so a mount reads EINVAL A line over 1024 bytes or with a control character was refused with its errno EINVAL but words 9ns reads as EIO (no `invalid` in them), and logged `err - exec`. They now say `invalid command line: ...`, and the root's err record names the pane the line would have run at. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* `@p<terminal>:N` past the terminal's last line is a miss, as in a fileGabriel Schneider4 days
| | | | | | | | The line-past-end check read only a file pane's lines, so a terminal's look went nowhere and said nothing. A terminal's lines are its logical lines, as a +Search of it lists them. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Tty in a missing directory fails and makes no pane; a shell that never ↵Gabriel Schneider4 days
| | | | | | | | | | | | | started answers runs The host could not start the shell (its chdir failed), and the pane stayed with no pty: a run written to it waited for a prompt for ever. Tty now refuses a directory that is not there, `Tty: <dir>: no such directory`, before making a pane; and a terminal whose shell the host could not start (Pardes.shellFailed, from each host's spawn) answers any run with `error shell gone`. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A Save refused for want of a name counts as a failure, so its announcement goesGabriel Schneider4 days
| | | | | | | | A builtin that would open a prompt nobody can answer (Save with no name on a scratch, through a ctl write) failed the write without saying a failure, so the announcement rule missed it and `msg N Save` stayed. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A root exec or look that fails logs its err against the pane it ran atGabriel Schneider4 days
| | | | | | | | `err - exec: ...` said nothing of where the line ran; the root's exec and look run at the keyboard's pane, and their err now names it, as a click there would. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Mount dials its peer at once and fails if none answers; a peer out of reach ↵Gabriel Schneider4 days
| | | | | | | | | | | | | is said as `dial failed: <reason>` Mount took any well-formed dial and nothing was found out until the first look through it, which then said `look: look: dial`. Mount now probes the peer (version, attach, its root) with the turn out, and fails its write, mounting nothing, when none answers; the dial errors read as words, `no answer`, `timed out`, `hung up`, and a look through a peer gone since names its path: `look: /n/peer/f: dial failed: no answer`. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A look that goes to another pane leaves the looking pane's dot aloneGabriel Schneider4 days
| | | | | | | | | | Every look that found something cleared the looker's selection first, so `@p5:2` or `f.txt:2` written to a pane's look collapsed its dot to q1-1. Only a look that stays in the pane starts from a clean selection now; one that goes elsewhere takes away just the click's highlight, and a miss changes nothing, as acme's look does. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* An event record never carries flag 2, as no expansion record follows; flag 1 ↵Gabriel Schneider4 days
| | | | | | | | | | | | reads the first word In acme flag 2 promises a second record with the expanded range, and acme's libraries read one. F records set it always and a click with no range known set it, with none following: a library then took the next event for it. Records here carry the word already, so 2 is never set. Flag 1 now reads the first word as acme's lookup does, so `Msg hi` is 1. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* `size` is refused only when the panes' own minima do not fit, so it is monotonicGabriel Schneider4 days
| | | | | | | | | | | | The check read how the weights shared the rows after the resize, not whether the rows could hold every pane's tag and two rows: a pane the proportional squeeze took short made `size` refuse, growing included, and refuse sizes it took a moment before. It now asks layout.minRows, the minimum placement asks, per column; a size they fit is taken, and a pane the resize took under its minimum gets its rows back. An out of range size says `invalid size`, which 9ns reads as EINVAL, as the doc had. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A look line takes `?re?` after its colon, as addr doesGabriel Schneider4 days
| | | | | | | | `file:?foo?`, `@p5:?foo?` and `:?foo?` were not taken for addresses (the first character had to be one of `0-9+-/$.#,;`), so they went the path way and found nothing. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* `@p<terminal>:<address>` takes any address over the terminal's logical linesGabriel Schneider4 days
| | | | | | | | | Only a line number reached a terminal (the line way); any other address went to lookAddress, which wanted a file and said the terminal was no text. The body is now the lines a +Search of it lists, addressed from the cursor, and the match selected as a +Search row's range is. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A line address to no open pane, `@p77:3`, is a miss said as one, not the ↵Gabriel Schneider4 days
| | | | | | | | | active pane `@p<serial>:<line>` goes the line way (parsePathLine), where a serial naming no pane quietly did nothing and look read back the active pane. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A builtin's Verbose announcement is logged only when it succeedsGabriel Schneider4 days
| | | | | | | | | | | A failure said while it ran (Exit's refusal, a Del refused) or later by the host (a Save's write, a Dump's) left `msg N Save` in the log as if it had worked. The announcement's record is remembered and taken back out on either, unless a follower read it already. A refused pane/new logs its err alone, and a failed Dump says its reason in words (`file not found`, not `FileNotFound`) so its msg goes and the err is the one record. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Every builtin says what it does in /commands, and a test holds each to itGabriel Schneider4 days
| | | | | | | | | 87 of the listed words had no description: settings have no doc comment to read, so each now carries a one-sentence doc, and the builtins without one got theirs. Exit's no longer names an internal function, and Repl's first sentence is short enough to be read whole. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* @p<serial>:<address> takes any sam address, as file:<address> doesGabriel Schneider4 days
| | | | | | | | | | @p<serial> took a line spot but an address after it (/re/, #n, $, 0/re/, a range) was dropped, the name failing the file check, and an unknown serial fell through to a search. The pane is now addressed like a file; a miss and a serial no pane has each log an err, and look reads back empty. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A word written to a raw terminal's look is listed, as in normal modeGabriel Schneider4 days
| | | | | | | | | A new Tty is in raw mode, where a right click is its program's, and a look written to it over 9P did nothing for that reason. The look file is no keyboard input: over 9P a plain word is listed in a +Search pane and read back whatever the terminal's key mode. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* One rule for failing builtins: the write fails, one err, no msgGabriel Schneider4 days
| | | | | | | | | | | | | | | | A builtin that failed through a ctl failed the write and logged its err, but through look, exec, tagexec or a column's exec it only said so on the message row, logged as a msg, and the write succeeded. Now every click write runs its builtin as a ctl line does (ctl.captured): a failure fails the write, with its words (EINVAL for malformed input, else EIO or what the words name), logs one err and no msg or announcement. get's Modified and a look miss are shown but logged once, as their err. A control character's refusal names its reason, a failed click reads back nothing, and look never reads back a pane closed since. The special case for a cut-short Edit through exec goes (the rule covers it); fs.md's table of exceptions becomes the one rule, as does the skill. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* get and a deleted file say what happened: the path, changed <serial> ↵Gabriel Schneider4 days
| | | | | | | | | | | | | | deleted, reloaded get on a file gone from disk failed with a bare ENOENT; it now says get: <path>: no such file. A watched file deleted on disk was silent; the pane keeps its text and says <file> deleted on disk (Save writes it again) once, logged changed <serial> deleted (only a path the watch has read: a pane renamed to a name never on disk is not deleted). A get that throws a dirty buffer's state away logs changed <serial> reloaded, as the watch's reload does. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A look that makes a +Search pane reads back that paneGabriel Schneider4 days
| | | | | | | | With LookWord list, or in a terminal, a look lists every place in a +Search pane, yet look read back the source pane, as if the hit were selected there. It now reads back the +Search pane it made. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A prompt a 9P request opens is an ask, answered with answer, never stackedGabriel Schneider4 days
| | | | | | | | | | | Save through exec on a terminal (or a scratch) opened its path prompt on screen, where no script could see or answer it, and a second Save stacked another. Opened while a 9P request is served, it now logs ask <serial> save path; answer <path> on the pane's ctl saves there and closes it, answer - closes it; asked again, the open one stands. A pane closed with it open logs answer <serial> -. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Newcol at the column cap says so, and a desktop session holds 16 columnsGabriel Schneider4 days
| | | | | | | | | | | Newcol with every column slot taken, or a column too narrow to split, did nothing and said nothing. It now fails (a ctl write with it) with no space for a column: 16 max (ENOSPC through 9ns), or that the column is too narrow. MAX_COLS is 16 on a desktop, 6 on the board; each column costs about 8 KB of core state (the Pardes struct grows from 1.47 MB to 1.56 MB), most of it its tag's text. fs.md and the skill give the cap. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* /commands takes each word's own doc comment, whole, without citationsGabriel Schneider4 days
| | | | | | | | | | | Repl showed Joincol's first line and Joincol another's, because a comment had drifted above the wrong declaration in builtins.zig; Kill's was cut mid-sentence and Exit's cited exec.c:892. The drifted line is back on Joincol, a parenthesis citing source (.c, .zig) is dropped, and the first sentence is kept whole up to 240 bytes, then cut between words with an ellipsis. A test checks every description opens its own word's comment. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* /commands says what each word does, and that the root's ctl takes EditGabriel Schneider4 days
| | | | | | | | | | | A /commands line gave a word, its argument and root or pane, but not what it does, and called Edit a pane's word though the root's ctl takes it too. Each line now ends with -- and the first sentence of the word's doc comment (read from builtins.zig itself, so there is one copy), and Edit is both. fs.md says a pane's word such as Undo or Msg is refused at the root. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A pane has a tagexec: a word run as a click in its tagGabriel Schneider4 days
| | | | | | | | | | The workspace had /tagexec and a column its exec, but a pane's own tag could be clicked only by an event record. pane/<n>/tagexec runs each line as a click in that pane's tag (never redirected to its event reader, as a header's is not) and reads back what it touched. The README, fs.md and the skill list it. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Refusals and records say more: get quotes its line, a bare answer lists the ↵Gabriel Schneider4 days
| | | | | | | | | | | | choices, a clean reload is logged get's refusal now quotes the line as a builtin's (Save's) does; a bare answer while a question stands says what it takes (answer takes k j or -); and a clean buffer reloaded from disk logs changed <serial> reloaded, as the kept dirty case logs changed <serial>. (A failed Save already says why: Save <path>: <the error in words>.) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A looked word selects its next place in its own pane, as acme's look3Gabriel Schneider4 days
| | | | | | | | | | | | | | A plain word looked at (a look write, a right click, an ML record written back) opened another identically named +Search pane each time, until the panes ran out, and never wrapped. As acme's look3 searches the text for it (look.c:210), it now selects the word's next place in that pane after the dot, wrapping at the end, and opens nothing. LookWord list, a new root setting, keeps the +Search listing (search is the default); the rsearch snapshot and an output test that walk the listing say LookWord list. (The leader-path check over every builtin gets a comptime branch quota, as one more builtin took the GUI build past the default.) Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A bare Shell names the shell it goes back to, and 12:0 says columns count from 1Gabriel Schneider4 days
| | | | | | | | | A bare Shell reset to the default without saying which; it now says Shell: the default again, <shell> ($SHELL, else /bin/sh). A line:col of column 0 said only address out of range; it now adds that a column counts from 1. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A column's ctl refuses as the other ctls do, and a bare answer says what it ↵Gabriel Schneider4 days
| | | | | | | | | | | | takes A column's ctl refused an unknown word in words 9ns maps to EIO, and took New extra as New. It now refuses the word as unknown control message (EINVAL, as the root and pane ctls) and any word but Tty with an argument as wrong #args. A bare answer on a pane's ctl says answer takes a choice or -, not unknown control message. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* A file changed on disk never reloads over unsaved editsGabriel Schneider4 days
| | | | | | | | | | | | | | | A watched file rewritten on disk was loaded over a buffer's unsaved edits and its dirty flag cleared, so Exit and Del no longer asked about them. A clean buffer still reloads; one with edits keeps its text and stays dirty, says <file> changed on disk (get reloads it, Save overwrites it), logs changed <serial>, and its Save warns once before overwriting the newer file, as acme's Put refuses a file modified since read (exec.c:577); get asks as ever (winclean, exec.c:513). The restored-watch test's dirty case now keeps the snapshot, and the filewatch snapshot script (re-recorded by name) checks both halves: the edited pane keeps its edit and says so, the clean one reloads. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* An Edit text block is found in linear time: only a . line is parsed againGabriel Schneider4 days
| | | | | | | | | | | Finding where an Edit block ends parsed the whole block again for every line, copying it into the scratch arena each time: a 10k-line $a block took minutes and tens of GB. sam_edit.waitsFor says what the block waits for, and while that is a text block's . line, only such a line is parsed again (editEnd, shared by Messages and completeEnd), its copy freed. A 10k-line block now takes well under a second and 50 MB of scratch. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* unlock's refusal says pane, not windowGabriel Schneider4 days
| | | | | | | "window not locked by this open" was acme's word; pardes has panes. Its test compares the constant, so it follows. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Repl's unknown-language refusal names a few whole and points to the restGabriel Schneider4 days
| | | | | | | | | The refusal listed every language and was clipped to fit, ending mid-word. It now names six, those that start as the word did first (pithon: python), says how many more, and points at docs/tags.md, which lists them all. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* size says its limits, keeps every pane its rows, and ptys follow their panesGabriel Schneider4 days
| | | | | | | | | | | | size refused an out-of-range line without saying the range, took a size too small for the panes (leaving bodies of no rows), and a terminal squeezed to its tag kept a pty of its old size. The refusal now names 20x6 to 4096x4096 (a bare size is EINVAL like any malformed line), a size that would leave any pane less than its tag and 2 rows is refused and undone, and a terminal's pty is resized with its pane down to one row (only a folded or unplaced one keeps its grid). Co-Authored-By: Claude Opus 5.5 <[email protected]>
* Tty and Shell refuse a directory: not a shellGabriel Schneider4 days
| | | | | | | | | | 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]>
* A detached session with no frontend takes its size from the root ctlGabriel Schneider4 days
| | | | | | | | | | A --detach session nobody is attached to stayed 80x24, too small for an agent's panes, and nothing could change it. The root ctl takes size <cols> <rows> while no frontend is attached (refused while one owns the size), and a client that attaches later still sets its own (the server compares with the core's size). Documented with the 80x24 default. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* +Errors with no room anywhere logs its text rather than fail the EditGabriel Schneider4 days
| | | | | | | | | Edit's p and = put their output in +Errors, and when its column had no room the placement refusal failed the whole Edit. +Errors now tries every column with room, the last first, and with none its text is logged as msg records, a line each; the Edit succeeds. Co-Authored-By: Claude Opus 5.5 <[email protected]>
generated by cgit v1.3 (git 2.53.0) at 2026-10-04 20:36:24 +0000