This implements basic parsing of sixel data. Lots of limitations and
temporary solutions as this is still work-in-progress:
* Maximum image size hardcoded to 800x800
* No HLS color format support
* Image is always rendered at 0x0 in the terminal
The 'last-row' variable points to the last row the cursor is *on*,
thus it's not a counter, and we need to add one when calculating the
new grid offsets, or we get an off-by-one error.
With this, there's no longer any need to scroll the reflowed text.
To do text reflow, we only need to know if a line has been explicitly
linebreaked or not. If not, that means it wrapped, and that we
should *not* insert a linebreak when reflowing text.
When reflowing text, when reaching the end of a row in the old grid,
only insert a linebreak in the new grid if the old row had been
explicitly linebreaked.
Furthermore, when reflowing text and wrapping a row in the new grid,
mark the previous row as linebreaked if either the last cell was
(the last column in the last row) empty, or the current cell (the
first column in the new row) is empty. If both are non-empty, then we
assume a linewrap.
A window resize doesn't necessarily mean the grid layout changes. When
it doesn't, don't reflow text.
This fixes a slowness issue on Sway, when swapping positions of two
~equally sized windows.
This was intended to prevent a new prompt from being printed on a
newline. However, it breaks normal output under certain conditions,
that have yet to be determined exactly what they are.
When current view is at a grid wrap around (last emitted row index is
< grid offset), the cursor row ended up being negative which we then
mapped to the top line.
This is wrong. When we're at a wrap around, re-adjust cursor by adding
the grid's row count.
To handle text reflow correctly when a line has a printable character
in the last column, but was still line breaked, we need to track the
fact that the slave inserted a line break here.
Otherwise, when the window width is increased, we'll end up pulling up
the next line, when we really should have inserted a line break.
The algorithm is as follows:
Start at the beginning of the scrollback. That is, at the oldest
emitted lines. This is done by taking the current offset, and adding
the number of (old) screen rows, and then iterating until we find the
first allocated line.
Next, we iterate the entire old grid. At the beginning, we allocate a
line for the new grid, and setup a global pointer for that line, and
the current cell index.
For each line in the old grid, iterate its cells. Copy the the cells
over to the new line. Whenever the new line reaches its maximum number
of columns, we line break it by increasing the current row index and
allocating a new row (if necessary - we may be overwriting old
scrollback if the new grid is smaller than the old grid).
Whenever we reach the end of a line of the old grid, we insert a line
break in the new grid's line too **if** the last cell in the old line
was empty. If it was **not** empty, we **don't** line break the new
line.
Furthermore, empty cells in general need special consideration. A line
ending with a string of empty cells doesn't have to be copied the new
line. And more importantly, should **not** increase the new line's
cell index (which may cause line breaks, which is incorrect).
However, if a string of empty cells is followed by non empty cells, we
need to copy all the preceding empty cells to the line too.
When the entire scrollback history has been reflowed, we need to
figure out the new grid's offset.
This is done by trying to put the **last** emitted line at the bottom
of the screen. I.e. the new offset is typically "last_line_idx -
term->rows". However, we need to handle empty lines. So, after
subtracting the number of screen rows, we _increase_ the offset until
we see a non-empty line. This ensures we handle grid's that doesn't
fill an entire screen.
Finally, we need to re-position the cursor. This is done by trying to
place the cursor **at** (_not_ after) the last emitted line. We keep
the current cursor column as is (but possibly truncated, if the grid's
width decreased).
Previously when updating a selection, we would unmark *all* cells in
the old selection, and then mark all cells in the new selection.
This caused *all* cells to be dirtied and thus re-rendered.
Avoid this, by adding a temporary state to the cells' selected state.
Before unmarking the old selection, pre-mark the new selection using a
temporary state.
When unmarking the old selection, ignore cells in this temporary state.
When marking the new selection, ignore cells in this temporary
state (except clearing the temporary state).
Maintain a view 'offset' (which glyph from the search string to start
rendering at).
This defines the start of the viewable area. The end is the offset +
the search box size (which is limited to the window size).
Adjust this offset whenever the cursor moves outside the viewable
area. For now, this is always done in the same way: set the offset to
the cursor position.
This means that when we're entering text at the end of the search
criteria (i.e. the normal case; we're simply typing), and the search
box reaches the window size, the cursor will jump to the start of the
search box, which will be empty. This could be confusing, but let's go
with for now.
This way, we don't have to manually insert flushes in code paths that
may execute outside of a wl_display_dispatch_pending().
(Those that execute inside a wl_display_dispatch_pending() are subject
to the flush performed at the end of the normal wayland FDM handler).
There are now three hook priorities: low, normal, high.
High priority hooks are executed *first*. Normal next, and last the
low priority hooks.
The renderer's terminal refresh hook now runs as a normal priority
hook.
With a bad behaving client (e.g. 'less' with mouse support enabled),
we can end up with a *lot* of xcursor updates (so much, that we
flooded the wayland socket before we implemented a blocking
wayl_flush()).
Since there's little point in updating the cursor more than once per
frame interval, use frame callbacks to throttle the updates.
This works more or lesslike normal terminal refreshes:
render_xcursor_set() stores the last terminal (window) that had (and
updated) the cursor.
The renderer's FDM hook checks if we have such a pending terminal set,
and if so, tries to refresh the cursor.
This is done by first checking if we're already waiting for a callback
from a previous cursor update, and if so we do nothing; the callback
will update the cursor for the next frame. If we're *not* already
waiting for a callback, we update the cursor immediately.
Since it doesn't block, we need to detect EAGAIN failures and ensure
we actually flush everything.
If we don't, we sooner or later end up in a wayland client library
call that aborts due to the socket buffer being full.
Ideally, we'd simply enable POLLOUT in the FDM. However, we cannot
write *anything* to the wayland socket until we've actually managed to
send everything. This means enabling POLLOUT in the FDM wont work
since we may (*will*) end up trying to write more data to it before
we've flushed it.
So, add a wrapper function, wayl_flush(), that acts as a blocking
variant of wl_display_flush(), by detecting EAGAIN failiures and
calling poll() itself, on the wayland socket only, until all data has
been sent.
In some cases, we end up calling render_refresh() multiple times in
the same FDM iteration. This means will render the first update
immediately, and then set the 'pending' flag, causing the updated
content to be rendered in the next frame.
This can cause flicker, or flashes, since we're presenting one or more
intermediate frames until the final content is shown.
Not to mention that it is inefficient to render multiple frames like
this.
Fix by:
* render_refresh() only sets a flag in the terminal
* install an FDM hook; this hook loops all terminals and executes what
render_refresh() _used_ to do (that is, render immediately if we're
not waiting for a frame callback, otherwise set 'pending' flag). for
all terminals that have the 'refresh_needed' flag set.
Instead of having the renderer calculate, for each cell, whether that
cell is currently selected or not, make selection_update() mark/unmark
the selected cells.
The renderer now only has to look at the cells' 'selected'
attribute. This makes the renderer both smaller and faster.
