| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| |
|
|
|
| |
Requires better support for special variable names in scenarios like
'screen' and 'console'.
|
| | |
|
| |
|
|
| |
..when building until layer 41
|
| |
|
|
| |
For starters start making the test fail when building until layer 41.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
It was stale from before I added the § sentinel.
|
| | |
|
| |
|
|
|
| |
More bugs about pointing outside text.
Still not done.
|
| |
|
|
|
|
|
|
| |
Time to turn on optimizations, since we aren't recompiling mu all the
time anymore.
But it doesn't help much with the editor. We need to be smarter about
not rendering the whole screen.
|
| |
|
|
|
|
|
|
|
|
| |
It can at most be just after the text of a line. If I click further
right, adjust it so it's at the end. If I click below, adjust it so it's
at the end.
Still not perfect. We aren't testing the value of the before-cursor
pointer, and it's actually still wrong when clicking below all the text
in the editor.
|
| |
|
|
|
|
| |
Another case I'm not writing tests for, after the
hide-screen/show-screen pair. But this case at least will eventually get
tests.
|
| |
|
|
| |
Had to take control of tb_present() after all. Termbox was wise.
|
| |
|
|
|
| |
But it doesn't work, because moving the cursor inside 'render' also
starts displaying it again.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
'render' now updates editor-data.cursor.
|
| | |
|
| |
|
|
| |
Wait, it works now?
|
| |
|
|
|
|
| |
Snapshot in switching editor-data.cursor to editor-data.before-cursor.
But I have trouble coercing events to touch events, even though using
the integer tag 2 for the conversion works.
|
| |
|
|
|
|
|
|
| |
Another bug in manually running the editor. At least show the error
message when you raise warnings in console mode.
Later we'll want to create a separate side channel and transparently
plumb warnings to the 'menu bar' of the editor..
|
| |
|
|
|
| |
Bug in manually running editor. Why wasn't it uncovered in the test?
Because I'm not yet rendering again from the event loop.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
Old plan: deduce cursor-row/column based on cursor.
New plan: deduce cursor based on cursor-row/column.
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
Html is a little more readable thanks to feedback from J David Eisenberg
(https://news.ycombinator.com/item?id=9766330), in particular the
suggestion to use https://addons.mozilla.org/En-us/firefox/addon/wcag-contrast-checker.
|
| |
|
|
|
|
|
|
| |
This takes away the need for multi-level escaping in scenario setup.
Post-conditions like screen-should-contain are still an issue, though.
Maybe each string literal should be able to specify escapes inside its
syntax? Or at a per-instruction granularity?
|
| | |
|
| | |
|
| |
|
|
|
|
| |
First, undo our existing hacky pseudo-support.
Some repl tests currently breaking.
|
| |
|
|
| |
But we certainly can't use it in the editor.
|
| | |
|
| |
|
|
|
|
| |
I still want to bring back replace-in-keyboard. No, not exactly that.
But saying press 8, press 8, press 8, over and over again, that's a
pain.
|
| | |
|