summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAge
...
| * | Insert Ctrl-s commits an undo checkpoint mid-sessionGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | helix commit_undo_checkpoint: u after the insert session goes back to the text as it was at Ctrl-s, a second u to before the session. It pushes an undo snapshot where the session stands. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Alt-u and Alt-U undo and redoGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | helix's earlier/later walk its history tree by time; pardes's history is a line, where earlier and later are undo and redo, so the keys are aliases. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | g. goes to where the last edit endedGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | helix goto_last_modification: every range moves to the end of the last change (the start of a deletion, the end of an insertion), extending in select mode. The position is kept on the Text in surface rows, set where setEditText hands a body its new text, so a terminal's typed text has one too and moves with its scrollback. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Alt-o, Alt-i and the other tree-sitter keys walk the file's syntax treeGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | helix's expand/shrink, previous/next sibling, all siblings, all children and parent-node end/start (Alt-o/up, Alt-i/down, Alt-p/left, Alt-right, Alt-a, Alt-I, Alt-e, Alt-b; Alt-n stays the new terminal) walk a whole-file parse the file keeps (File.node_tree), made on the first such key after an edit and reused until the next, never per key. syntax.walkNodes follows helix's TreeCursor over every node, anonymous ones included, and Alt-i goes back to what Alt-o grew from while the selection still holds it. The harness runs without grammars, so the cases prove the plain-text no-op; a test replays eleven JSON walks whose results were taken from the installed hx. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | K and Alt-K keep or drop the ranges a regex matches in; Alt-: turns every ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | range forward K and Alt-K share the s/S prompt, which now carries a mode rather than a split flag: a range stays when a match starts inside it (K) or when none does (Alt-K), the primary becomes the first, and keeping none leaves the selection as it was (helix keep_or_remove_matches). Alt-: is ensure_selections_forward. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | A selection holds up to 1024 ranges on the desktop, 64 on the boardGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The limit was 64 everywhere, which helix-golf's enumerate_and_align outgrows (%s |\d+ makes 121 ranges). It is now memory.limits.selections, per platform, and the counts that hold it are u16. The ranges stay inline in each Text, so a pane grows by about 80 KiB on the desktop (three texts and the regex prompt's snapshot); the Pane size budget in Terminal.zig moves from 128 to 160 KiB and says why. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Q records keys into a register and q types them again; . repeats the last insertGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Macro.zig keeps keys to be typed again. Q records into @ (or "<reg>) in helix's key notation until the next Q; q replays a register's keys count times through the modal handling, where Esc does not hop panes and Enter, Tab and Space neither look, execute nor open the leader. . replays the normal command that entered the last insert session once and what was typed in it count times (helix repeat_last_insert). a at the end of the text now adds the newline helix's append_mode does, which a terminal's text does not. helix-golf's invert_dictionary passes. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | $ keeps the selections a shell command exits 0 on, and line end is glGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | $ was a second key for line end, which helix never meant: in helix it is shell_keep_pipe. It now arms the pipe prompt with a $ marker, runs the command once per range with that range's text on stdin, and keeps the ranges it exits 0 on; the primary stays if kept, else the last kept range takes over, and keeping none changes nothing. The runner's answer is all or nothing, so the command runs in a subshell whose status is echoed as its output. Line end stays on helix's own gl and End; the snapshots, tutor and docs that pressed $ for it press gl. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | A tag's edited text stays alive until the next edit, so ms and md in a tag ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | read it whole setEditText freed the text a tag edit was handed once it had copied the tag's own part out of it, but every caller reads that text back to place the cursor, and surround goes on to edit it again: ms and md in a tag read freed memory. The same held for a terminal without an edit buffer and a text with no pane. The text is now kept until the next such edit (edit.retireEditText), which covers every caller at once; the callers the audit found reading it back are surround's two, the paste and replace paths, the number adjust, and the insert-mode Enter. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | The helix-golf steps that walk search hits are waived, so hxgolf can passGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | reverse_golf_example from ""N on needs *, N and n to walk the search register; in pardes those keys are the acme look ring, by decision. The seven steps are pinned in golf-waivers.jsonl, so a new mismatch stands out. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Every range lives through insert mode, carried through each editGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Insert mode dropped the selection on the first key and restored only the primary's appended-over span on Esc. Now each insert-mode edit is made at the bare cursor as before, and the range is then carried through it with helix's Range::map rules: typing slides an i range and stretches an a range, Enter (which now also drops the blanks before the cursor, as helix does) slides or stretches by what the cursor moved, and an arrow collapses. The replay gives every other range the same. Esc after a pulls each range's end back by one (helix restore_cursor), which replaces Text.append_at. The wiX-edit-drops-sel and msel-append waivers are gone. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Registers hold a value per range, and "<reg> names oneGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Registers.zig replaces the one yank buffer: every register keeps a value per range, y fills them in document order, and p, P, R and insert Ctrl-r put value i at range i, repeating the last (helix paste_impl). "<reg> names the register for the next command; _ swallows, # numbers the ranges, . is each range's text, % the file's name, / the last s/S pattern, and + and * are the system clipboard through the ClipYank and ClipPaste paths. SPC y now writes + alone, as helix's does. The acme chords and a paste into a terminal take the default register joined by newlines. The msel-yank-paste waiver is gone. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | The two regex waivers are gone: s and S already match helix thereGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | s and S search each line as its own haystack, so ^ and $ hold at every line and . never takes a newline, which is helix's multi_line regex. The waivers for the old difference, sel-regex-caret and sel-regex-dot-newline, had gone stale, and --strict stopped on the first before printing its summary. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | docs/selections.md says how normal mode holds its selectionsGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | The primary and the others, cells against helix's gaps, how setRanges normalizes, which commands replay per range and which make one edit, and what undo, the register and repeat do with several ranges, with the places helix differs listed at the end. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | & aligns the selections into columnsGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | helix-golf's enumerate_and_align ends on %s |\d+<ret>&. align_selections treats the k-th range of each line as column k and puts spaces before each range until its head reaches the widest head of its column, measured in display cells, which File.rawDisplayCol already gives. A range over several lines refuses, as in helix. The example itself still needs more than 64 ranges, which the selection model does not hold. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Backspace in a line's leading blanks goes back to the previous indent stopGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | helix's delete_char_backward dedents: when everything before the cursor on its line is blank, it deletes back to the previous multiple of the indent width, a whole unit when already on one (a tab before the cursor still goes alone). pardes deleted one space, so O<backspace> under an indented line left three of four, which is where helix-golf's function_into_class ended up. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Alt-( and Alt-) rotate the text of the selectionsGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | helix-golf swaps keys and values with rotate_selection_contents (webS_to_ <ret><A-(>, then per line). Each range's text moves to the next range (the previous one for Alt-(), in one edit; a count rotates by that many ranges, each range comes back over the text it now holds in its own direction, and the primary follows its text. That is the harness's helix; 25.07.1 read the count as a group size and left the primary alone. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Alt-J joins lines and selects the spaces it put inGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | helix-golf's text_into_array turns lines into an array with %<A-s>ms"<A-J>i,: join_selections_space joins like J and leaves one cursor on each space it inserted. J and Alt-J now make one edit over every range, as helix does, rather than J being replayed per range; that is what lets Alt-J hand back all the spaces as the new selection. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Blank lines open above a selection's first line and below its lastGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | [<space> and ]<space> measured from the cursor's line, so over a selection of several lines whose head was at the bottom, [<space> opened a line INSIDE the selection. helix takes the range's line_range: above its start, below its end. helix-golf's export_from_mod (yp[<space>) is where it showed. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Replacing a selection with the register keeps its directionGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | R always left the head on the pasted text's last char. helix maps the replaced range through the change, so a backward selection comes back backward, its head on the first char: helix-golf's reverse_golf_example (xbR) is where it showed. SPC R shares the path. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Surround keeps the direction of the selection it wrapsGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | ms<ch> always left the wrapped range pointing forward. helix gives each range back with_direction(range.direction()), so a selection made backward (b, or v moving up) keeps its head on the open, which is where the next motion starts from. helix-golf's csv_to_sql (bms`) is where it showed. Co-Authored-By: Claude Opus 5.5 <[email protected]>
| * | Every helix-golf example is a differential case, one per step of its walkthroughGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The ten examples on helix-golf are imported as 118 cases: each numbered step of an example's walkthrough is the case whose keys are the command up to and including that step, so the first mismatch names the step where pardes and helix part. Six more restart at a step that begins with %, from helix's own text there, so a later step is tested even when an earlier one misses. zig build hxgolf compares all 124 with goldens from hx-harness. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | The board build compiles again: pty status and run read no prompt marks ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | where there are no terminals pty.zig reached into a terminal's stream handler, which is void on esp32p4, so the board build stopped at readStatus and writeRun. Both now know at comptime that such a build has no marks. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | The docs say which shell a command runs in, what Kill's word matches, where ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | dumps go and in what time, and how to send code to a REPL Dogfood found the docs saying $SHELL -c where commands run the root ctl's Shell, the README's log line missing run, exit, send, dump and restore, Kill's argument unexplained, and nothing on Repl - and bare Repl, on bracketed paste for multi-line code over pty/data, or on the UTC dump time. The Restore write's ECONNRESET through a mount is documented rather than fixed: pardes sends the answer before the hang-up, and 9ns drops it when its next send fails first. The REPL decision is recorded with the others, and the fs guide says a builtin's word still runs first. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A REPL question's answer goes to the pane its letter offered, or says that ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | one is gone The letter was looked up again when answered, so a REPL closed or rebound while the question stood could hand the text to another pane with the same letter, or drop it without a word. The question now keeps each letter's serial; the answer is that pane, still bound and running, or a message. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | Errors are said in words, and a refusal names what it takesGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | A failure read `Del k|j: InvalidArgument`, `Kill: NoSuchCommand`, `ill-formed control message`, or nothing at all. reportError now spells an error's name as words, and the refusals name their accepted forms: pty/ctl lists its verbs, a flag file takes 1 or 0, a setting word takes on or off, Del takes k, j or nothing, Repl takes - or a language. Kill with nothing to stop says so, /focus says pane, a Look that found nothing says no match, and pty/run on a command pane says it is not a shell. A ctl write's failure has room for the Repl list and is cut between words, the Repl word at a character. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A REPL's language is named and found as the syntax table finds one: aliases, ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | any case Repl took only the grammar's exact name, and a file's language was found by a case-sensitive suffix of its path, apart from the aliases a code fence already had. The aliases move to grammar_manifest with byName and byPath, which syntax, Repl and langOf share, so `Repl py` and a .PY file work. The line-by-line warning is for the python language, not an id that starts with python. Joincol's doc line was sitting on Repl. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | Focus on a closed pane goes to the pane that takes its place, never to the ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | one focused before Dogfood closed a pane in one column and focus jumped to another column, only because that pane was focused last. Now removePane decides for every route (Del, a ctl del, a command pane's end, a 9P remove, a stillborn scratch replaced): the pane that grew into the rows, or the scratch standing in an emptied column; removeColumn gives it to the pane level with it in the column that takes the width. Closing an unfocused pane leaves focus alone. The per-caller focus lines in Del and replaceStillborn are gone, and so is prevFocus. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A build that runs no commands refuses a word that is not a builtin, and says soGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | Board and web have no spawn. An unknown word there used to fall through to the command path and do nothing visible; now it answers with a failure the caller sees. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | The web and board builds compile again: a shell's lookup and a dump's read ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | are the host's The Tty shell check sized a buffer with std.fs.max_path_bytes and the Restore check read the dump through the host filesystem, both in the core, which the wasm32-freestanding build cannot compile (no PATH_MAX, no libc). Both now run only where there is a host, as the rest of that code does. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | Text sent to a REPL is plain lines, sent whole or not at allGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A CRLF became two line ends, the blank line ending a Python block; an ESC in the text, ESC[201~ above all, could end a bracketed paste early; and on the board, whose writes are 128 effects of 64 bytes with nothing pending, a send past 8 KB was cut and lost its closing ESC[201~ and Enter, leaving the REPL in a paste. Lines now end in LF alone, ESC and every control byte but tab and newline are dropped, and a send larger than can be written at once (Pardes.writeRoom) is refused with a message instead of cut. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A command's exit is told once its output is in, by the pty's state, not a timerGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 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]>
* | | A click written back to event is done as the click, REPL and all; a REPL's ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | ctl line names it Writing an M record back from a .py body with a REPL bound ran the text as a shell command (exit 127), where the click itself went to the REPL. acme does what the click would have done (xfideventwrite, xfid.c:842-844), so a body X record written back now goes the way a middle click goes: to a bound REPL, bracketed or line by line, the several-REPLs question included. An F record, a line written to exec, was never a click and still runs as a command. For a readable list of bindings, the smallest place: a bound terminal's ctl line ends with its id after current/notcurrent. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | Whether an exec came from a body travels with it, so every selection goes to ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | the REPL The execute key and the middle click marked a global flag that the first execute cleared, so of several selections executed at once only the first reached a bound REPL and the rest ran as commands. Where an exec came from now goes with it, as acme's execute() asks of its Text (exec.c:122,169): builtins.Ctx carries from_body, runBuiltinFrom and executeFrom take it, the gestures set it, and chordEachSel sets it for each selection. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A detached session stopped by SIGTERM, SIGINT or SIGHUP unlinks its socketsGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The signals' default action ended the process where it stood, leaving pardes-9p-*.sock and pardes-detached-*.sock for the next session to trip on. The detached session now takes them to mean quit: the handler sets a flag and wakes the loop, which leaves as Exit does, through the teardown that unlinks both. fs.py sends each of the three and checks the sockets are gone. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A 9P request that fills a stillborn scratch's column replaces the scratch ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | before it is answered A pane that joins a column holding only an empty +New takes the column, but that replacement ran at the end of the editor's next update, not of a 9P request: a look answered 19 while index still listed the +New 18, whose del came later. serveFs, a step of its own, now runs it too before answering. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A restored terminal's tag names neither the REPL it was nor the shell it hadGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A dump keeps a terminal's whole tag, and restoring read its tail from it after the directory: the REPL id and Tty+fish, which were live facts and not the tag's text, came back as a custom tail, so the tags said python-a over a fresh shell bound to nothing. Restoring now sees through the id and the shell in the Tty word, a default tail stays the default, and the new shell names itself when it starts. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A REPL takes text only while its program has the terminalGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A binding outlived its program: a command pane reused kept the REPL of the command before, a done command pane could be bound, a document pane (a terminal with no emulator) was accepted, and after Ctrl-D in python3 the text went to the shell, which ran import os as ImageMagick's. A command pane's binding now ends with its command and a done one is refused, a pane with no emulator is refused, and an interactive terminal's REPL is live only while a program other than its shell holds the tty (the host's tty_taken); when none of the language's REPLs is, nothing is sent or run and the pane says so. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | Tty with a shell that is not there says so, and never sets the caller's shellGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Tty /nonexistent or Tty fsh started the host's fallback shell without a word, and when spawnTty made no pane, the argument became the calling pane's shell. Tty now looks the shell up first and fails 'Tty: no shell "fsh"' when it is neither a name on the usual paths nor a path to one; spawnTty answers the pane it made, the one given the shell, and says why when it made none. The lookup (host_io Shell.find) gives the turn up around its access() calls, since a typed path may be under a mount this editor serves, and forkShell copies the pane's shell before looking it up and checks the pane is still its own after. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A Restore of a dump that cannot be read fails at once, and spends no warningGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The dump was read only by the host after the step, where a failure went to the message row alone: a Restore written to ctl answered success, logged no err, and a second one past the unsaved-text warning did the same. Restore now reads the dump first, before the warning, and says 'Restore: <path>: no such file' (or permission denied, a directory), which fails a ctl write and so logs err. A bare Restore with no dump yet, or a name too long, says so too. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | The log's run record carries the whole command line, after the new record of ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | its pane run logged the line's first word only, and before the new record of the pane it ran in, since panes are announced at the end of a step. It now carries the whole line as msg carries its text (control characters, newlines too, as spaces), and announces the new pane first. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A command pane echoes its first command as it does the later onesGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | Only a reused pane showed '% line' above its command, so the first command's output stood with nothing saying what made it. Every command is echoed now, the first too. The goldens of chordcut, cmdexit, cmdword and layout-open gain that line and the rows below it move down one. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | A command of several lines keeps its pane's tag to one row, and the tag ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | never panics on a longer prefix A two-line exec made a command pane whose tag prefix held the line's newline, so the tag ran to a second row and tagTintRange sliced its first row past its end: panic. acme runs a multi-line selection as one command, and so does pardes; the tag now shows the line's first row, cut at a character's start, and '...', tagTintRange stays inside the row it is given, and the '% line' echo turns the line's newlines into the terminal's, so it does not staircase. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | | The 9P docs find the mount through $NINE_MOUNT, and then the editor is just ↵Gabriel Schneider31 hours
|/ / | | | | | | files
* | A terminal bound as a language's REPL takes the execs made on that ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | language's files Repl python in a terminal's tag or on its ctl binds it (Repl - unbinds, Repl bare says the binding, the same word again changes nothing), with an id, python-a, python-b, in its tag. An exec made by a gesture on the body of a file in that language -- a middle click or the execute key, on any text, a single word too -- or on the REPL's own body is typed into it, bracketed if its program asked for that, else line by line with a once-said warning for Python's blank lines, then Enter; the message row says -> python-a in the tag's name tint, the log send <from> <to> <id>. Several bound for the language: the pane asks which on its notice band, as Del asks, one key answering, Esc none, nothing remembered. The gesture is known where it is dispatched (mouse.zig, the execute key), so the tag's words, Exec <text> run by name and a command word @`cmd` stay commands, and a 9P exec is never sent. A finished command pane is no REPL; bindings are not dumped. The docs spell out the escapes, the pdb and input() caveat and the no-bracketed-paste papercut. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | Kill signals a command only while it runs, and only the pane it was meant forGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | A signal effect carried only a slot, so a Kill followed in the same step by the pane's closing and its slot's reuse signalled the new pane's child; and a command pane whose child had exited and been reaped kept a process group id another process could take. The effect now carries the pane's serial, checked as it is performed, and a host signals no command whose exit it has recorded; the core already sends Kill only to a command it has not been told is done. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | A command pane's command is over when its process exits, not when its pty closesGabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A job left in the background (sleep 100 &) held the pty open, so the pane stayed running and the child a zombie until the job ended; a command that closed its terminal and ran on got its end of file at once, the host waited 100 ms for an exit, reported exit ? and hung it up. Now each command's child is watched on a thread (waitid with WNOWAIT, so its pid stays its own until the host reaps it), and the host tells the core the exit from that: after the pty's end of file, so the output before the exit is in, or 50 ms after the exit without one, a job holding the pty. The pty stays open until both, so a command that let go of its terminal is never hung up by it. All four front ends; a host that cannot start the watcher reads the exit at end of file as before. Tests: host_io's for both cases, and cmdexit.snap end to end. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | A reused command pane leaves the last program's screen modes before the next ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | command A program that crashed out of a full screen left the command pane's emulator in the alternate screen with mouse reports, bracketed paste or a hidden cursor on, and the next command ran in that. Before the '% line' separator the pane now leaves the alternate screen, turns those modes off and soft-resets (DECSTR). The docs add that a command which clears the screen and its scrollback erases the pane's history above it. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | Only Tty reads Word+arg as its argumentGabriel Schneider31 hours
| | | | | | | | | | | | | | | | Every builtin that takes an argument split a word at its first +, so Dump+2024.zon ran Dump with 2024.zon. A builtin now opts in with plus_arg, read by the registry, and only Tty does. Co-Authored-By: Claude Opus 5.5 <[email protected]>
* | A spawn names the pane it was made for, so a pane closed and its slot ↵Gabriel Schneider31 hours
| | | | | | | | | | | | | | | | | | | | | | | | | | | | retaken spawns nothing The spawn effect carried only a slot, and the host read the command and shell off whatever pane held the slot as it forked: a command pane closed and its slot taken by the next one before the effects ran forked the new pane's command twice. The effect now carries the pane's serial and is dropped when the slot holds another; forkShell checks it again after giving up the turn for its directory's stat, the one moment a 9P client can change the panes under it. Every front end performs spawns with the turn, from its step. Co-Authored-By: Claude Opus 5.5 <[email protected]>
generated by cgit v1.3 (git 2.53.0) at 2026-10-02 10:09:32 +0000