summaryrefslogtreecommitdiff
path: root/test/snapshots/msg.snap
diff options
context:
space:
mode:
authorGabriel Schneider <[email protected]>2026-11-08 23:21:47 -0300
committerGabriel Schneider <[email protected]>2026-11-09 01:28:28 -0300
commit094b562b5b90558e82a9c3e2beea386d93bd1803 (patch)
tree23e49a2f2c308acbacba532a126bb6b1631d7aca /test/snapshots/msg.snap
parent80db18269dd9f7153a02f3efa6ff92c69948a623 (diff)
downloadpardes-094b562b5b90558e82a9c3e2beea386d93bd1803.tar.gz
pardes-094b562b5b90558e82a9c3e2beea386d93bd1803.zip
tests: the boot, argv and builtins snapshot scripts (22: argv badutf boot builtins dump load restore headersweep headersweep-restored logquiet tutor topbarexec leader mini ninep nested nested-optout shellset msg winch savepane filewatch) are process cases, test/process/tty_boot.zig, each its script's steps as Tty calls and its golden unchanged as its expectation; checked against the old runner by two breaks, the message stamp's spacing (3 fail in both) and the tty reading a middle press as a right one (the same 9 in both)
Co-Authored-By: Claude Fable 5.1 <[email protected]> Claude-Session: https://claude.ai/code/session_016Y9MNAtZenFv4weoMD3cge
Diffstat (limited to 'test/snapshots/msg.snap')
-rw-r--r--test/snapshots/msg.snap103
1 files changed, 0 insertions, 103 deletions
diff --git a/test/snapshots/msg.snap b/test/snapshots/msg.snap
deleted file mode 100644
index 64a77bb2..00000000
--- a/test/snapshots/msg.snap
+++ /dev/null
@@ -1,103 +0,0 @@
-# The transient message row: the LAST row of a pane, left-aligned at the text
-# area's origin, saying what just happened to that pane and when. An OVERLAY,
-# not geometry — nothing shrinks to make room, so a pane with nothing to say is
-# not touched at all, which is why the other 71 goldens only moved where a
-# prompt or a message actually appears.
-#
-# The clock reads `--:--:--` here and only here: $PARDES_NOTIME (set by
-# snapshot.zig, the same flag that hides the language backend's durations)
-# blanks the DIGITS rather than the row, because a message that disappeared
-# under the harness could not be pinned at all.
-#
-# The row has exactly two possible occupants and they are not the same kind of
-# thing. A MESSAGE is a report of something that already happened, and the next
-# key or mouse event wipes it. An armed PROMPT is what that keystroke is being
-# typed INTO, it lives until Enter or Esc, and it WINS.
-lines a.txt 6
-start 15 70 a.txt
-wait 8000 Newcol Joincol
-wait 8000 a.txt
-stable 700 20000
-snap opened
-# Save (SPC f s) writes the file; the SHELL is where the write and the clock
-# both are, so the shell is what posts the message once the bytes are down.
-key space f s
-stable 700 15000
-snap saved
-# A REPAINT is not input. A resize redraws every cell in the window from
-# scratch and the message is still there on the other side of it — the row is
-# state, not one frame's flash. (Different rows, or the kernel sends no
-# SIGWINCH and nothing repaints.)
-resize 17 70
-stable 700 15000
-snap repainted
-# ...but the next KEY takes it away. `j` is a plain cursor motion and changes
-# nothing else on screen, so the vanishing message is the whole diff.
-key j
-stable 700 15000
-snap keycleared
-# Again, and this time a bare mouse MOTION — no button, nothing clicked —
-# clears it: any mouse event counts, the rule is "you are back", not "you did
-# something".
-key space f s
-stable 700 15000
-snap saved-again
-motion 20 6
-stable 700 15000
-snap mousecleared
-# The other message: an external write lands in the pane (filewatch.snap owns
-# that contract) and the row says which of the two things happened. `run` is a
-# child of the RUNNER, so the app receives no keystroke here and nothing but
-# the watch could have posted this.
-run printf 'EXTERNAL\n' > a.txt
-wait 10000 EXTERNAL
-stable 700 15000
-snap reloaded
-# `/` still types into the tag tail — same buffer, same offsets, same one-line
-# editor, same Enter and Esc — but it is DRAWN here now, so the tagline keeps
-# its builtins instead of losing them to a long pattern. Note the tag row above
-# ends at `Del`: the marker and the pattern are no longer on it at all. (The
-# `/` keystroke also cleared the reload message, which is the rule above.)
-key s
-text EXT
-stable 700 15000
-snap prompt
-# ...and a message arriving UNDER an armed prompt does not disturb it: a second
-# external write reloads the pane and posts one, the row stays the prompt's.
-run printf 'AGAIN\n' > a.txt
-wait 10000 AGAIN
-stable 700 15000
-snap prompt-wins
-# Esc abandons the input, and the very key that abandoned it is also the key
-# that clears the message queued underneath — so the row goes empty rather than
-# handing back a report from before you started typing.
-key esc
-settle 100
-stable 700 15000
-snap esc
-# Every other writer of this row is something going WRONG, which is a poor way
-# to look at a row: `Msg <text>` is the word that writes it on purpose. An
-# ordinary builtin, so it runs the ordinary way — write it in the buffer,
-# select the line, execute it — and what lands is the text, verbatim, with the
-# same stamped-looking left edge as the reports above. (The clock belongs to
-# the SHELL, which is why this one has none: the core has no clock.)
-key o
-text Msg hand-written report
-key esc
-settle 100
-key x
-key tab
-stable 700 15000
-snap msgbuiltin
-# A message wider than the pane keeps its TAIL, not its head: the end of a
-# message is the part that says something -- a path's file name, an error's
-# reason -- so the head is what a narrow pane goes without.
-key j
-key o
-text Msg 0123456789 0123456789 0123456789 0123456789 0123456789 0123456789 0123456789 THE-END
-key esc
-settle 100
-key x
-key tab
-stable 700 15000
-snap msgtail