Previously ref-tests just ignored the scrollback history to keep
the old tests working, this would lead to new tests which rely on
scrollback history to succeeed even though they should not.
This has been fixed and it is now possible to create ref-tests with and
without scrollback history. When available the scrollback history is
compared, but the old tests still work without having to adjust them.
This fixes#1244.
In the current scrollback PR the `reset` command does not affect the
scrollback history. To make sure the terminal is properly reset, it
should clear the scrollback history.
To make resetting efficient, instead of resetting the history,
the scrollback history is hidden by setting `grid.scroll_limit` to `0`.
This will not clear the history but instead just make it inaccessible,
which should have the same effect.
The visible area is reset by the shell itself, so in combination this
clears the complete terminal grid from a user perspective.
This fixes#1242.
This bug was introduced by the commit which fixed the invisible cursor
in the first column (54b21b66ecc6f8f149d1425567e0e3d766a3ac54). To
resolve this the alternative implementation by @jwilm has been applied
which seems to work out.
This fixes#1259.
There was an issue where the lines would be messed up when the terminal
was resized, this was because lines were just added/removed at the end
of the buffer instead of the actual end of the terminal (since the end
of the terminal might be in the middle of the buffer).
This has been fixed by relying on `self.zero` to determine the position
of the start of the terminal and then calculating where lines have to be
inserted/removed.
Some tests have also been added with documentation that should make it
a little easier to understand how the process works and how the raw
buffer is layed out.
This should all work no matter how big the scrollback history and even
when the currenty viewport is not at the bottom of the terminal output.
There was an issue where alacritty would panic whenever the scrollback
history size is set to 0, this fixes that issue.
The panic was caused by a substraction with unsigned integers which was
underflowing, this has been fixed to use `saturating_sub`.
After that was fixed there was still a bug where scrollback would not
behave correctly because the number of lines in the grid was decided at
startup.
This has been adapted so whenever the size of the terminal changes, the
scrollback history and grid adapts to make sure the number of lines in
the terminal is always the number of visible lines plus the amount of
scrollback lines configured in the config file.
This fixes#1150.
There was a bug in the display iterator where the first column was never
reached after the top line because it was instantly incremented to 1
after it was reset when iterator column reached the end of the terminal
width.
This has been fixed by making sure that the column is never incremented
when the column is reset due to a change in terminal line.
This fixes#1198.
The latest selection changes broke a few tests, these have been
corrected.
Two of these tests were broken because they assumed different span
types, the test have been changed here because the result was correct.
One test did actually catch a bug where selection of two cells from
right to left would incorrectly mark the cells as selected even though
they should not have been, this has been fixed in the `simple_span`
method.
Some tests are still not passing, though.
A migration script was added to migrate serialized grids from
pre-scrollback to the current format. The script is included with this
commit for completeness, posterity, and as an example to be used in the
future.
A few tests in grid/tests.rs were removed due to becoming irrelevant.
This changes two things, the first thing it does is that now whenever a
keybinding sends an escape sequence, the viewport is automatically
scrolled to the bottom.
This is enabled by default and fixes#1187.
The second thing is automatic scrolling when a command writes to the
terminal. So when running a command like `sleep 3; ls -lah`, alacritty
will scroll to the bottom once the output is sent, even if the viewport
is currently not at the bottom of the scrollback.
Because this can have an impact on performance, and is not enabled by
default in terminals like iTerm or Termite (VTE), it is an opt-in
setting in the config.
The current `span_simple` selection is everything but simple. This
version should have the same functionality as the current `span_simple`
with the difference that a lot of complexity has been removed.
Not only is this code shorter, it should also be significantly easier to
understand with no "magic" to it. This will hopefully prevent us from
having an unmaintainable blob of off-by-one guessing in the repo.
Also removed the `out` file which I used in the original PR because
scrollback is not implemented yet. :)
The different scrolling methods added a bunch of boilerplate where the
call was just forwarded to the next struct, this has been removed by
making the scroll amount into a struct.
Now everything is called through one method and the parameter decides
how far the viewport should be scrolled.
This offers a few additional hotkeys that can be used in combination
with scrollback. None of these are used by default yet.
This implements the following bindings:
- ScrollPageUp: Scroll exactly one screen height up
- ScrollPageDown: Scroll exactly one screen height down
- ScrollToTop: Scroll as far up as possible
- ScrollToBottom: Scroll as far down as possible
This fixes#1151.
When selecting to the top and starting in the first cell, alacritty
would crash. These cases have been fixed and now selection should be
completely working.
When scrolling down with a selection on screen the first line was not
properly selected. This has been fixed by making sure the selection
always starts in the first cell when it is only partially visible.
When the user selected multiple lines, dragging the selection downwards,
and then leaves the cursor to the left side of the first cell, the first
cell was still incorrectly selected. This has been fixed.
The selection also did not update if the mouse was outside of the
window, now all movement events are accpeted even when the mouse is
outside of the window. This allows updating the selection when the user
is dragging the cursor too far.
Mouse movement and click events outside of the window are not
propagated, these are only used for updating the selection.
There were a few issues with selection in scrollback that were mainly
off-by-one errors. This aims at fixing these issues.
This also fixes a bug that currently exists in master where the last
cell is not selected when the mouse leaves the window to the right.
Because there was some overlap with branch #1095, these two PRs have
been added together and the config has been restructured to make use of
a `scrolling` section.
The default faux scrolling amount has also been changed to `3` because
this simplifies the code and falls in line with what most other terminal
emulators do.
There should be no additional test failures due to this.
There was an issue where alacritty tries to convert the lines in a
selection to the on-screen lines even when the selection is not on the
screen. This results in a crash.
To prevent this from happening the selection now is not shown if it is
off the screen.
There currently still is a bug that when the selection is at the top of
the screen but still half visible, it will not show the top line as
selected but start in the second line.
This bug should be resolved with
https://github.com/jwilm/alacritty/pull/1171.
This fixes#1148.
To make it possible to access the native scrollback buffer in the
alternate screen without having to disable faux scrolling, faux
scrolling is now disabled when the `shift` key is held down.
This should allow alacritty to have the best of both worlds, a native
scrollback buffer in the alternate screen buffer and faux scrolling.
When implementing fallback to the default value with an u32 you will get
0 as the default value.
However the default scrollback value is 10_000. A custom deserializer
has been implemented which automatically falls back to the correct
default value.
Selections now *mostly* work. They move as the buffer scrolls, copying
works as it should, and it looks like the different selection modes
behave properly as well.
The new Selection implementation uses buffer coordinates instead of
screen coordinates. This leads to doing a transform from mouse input to
update the selection, and back to screen coordinates when displaying the
selection. Scrolling the selection is fast because the grid is already
operating in buffer coordinates.
There are several bugs to address:
* A _partially_ visible selection will lead to a crash since the drawing
routine converts selection coordinates to screen coordinates. The
solution will be to clip the coordinates at draw time.
* A selection scrolling off the buffer in either direction leads to
indexing out-of-bounds. The solution again is to clip, but this needs
to be done within Selection::rotate by passing a max limit. It may
also need a return type to indicate that the selection is no longer
visible and should be discarded.
* A selection scrolling out of a logical scrolling region is not
clipped. A temporary and robust workaround is to simply discard the
selection in the case of scrolling in a region.
wip selections
fix issue with line selection
selection mostly working
need to support selection not being on the screen at draw time
Fix selection_to_string
Uncomment tests
Supporting selections with scrollback has two major components:
1. Grid needs access to Selection so that it may update the scroll
position as the terminal text changes.
2. Selection needs to be implemented in terms of buffer offsets -- NOT
lines -- and be updated when Storage is rotated.
This commit implements the first part.
This intends to optimize the case where the top of the scrolling region
is the top of the screen. In theory, scrolling in this case can be
optimized to shifting the start/end of the visible region, and then
rearranging any lines that were not supposed to be scrolled (at the
bottom of the region). However, this didn't produce quite the speedup I
expected.
In addition to a marginal performance improvement, this simplifies some
logic in the Term implementation since now the Grid fully handles row
recycling.
Christian Duerr
Christian Duerr
Joe Wilm