DECGRI, i.e. repeat sixel character, only need to do image resizing,
and updating the current ‘column’ value *once*.
By adding sixel_add_many(), and doing the size/resize checking there,
the performance of sixel_add() is made much simpler.
This boosts performance quite noticeably when the application is
emitting many and/or long repeat sequences.
This results in the same number of instructions inside the loop, with
a ‘lea’ instead of a ‘mov’, but simplifies the post-loop logic to
update the global state.
All-empty pixels rows in the last sixel row should not be included in
the final sixel image.
This allows applications to emit sixels whose height is not a multiple
of 6, and is how XTerm works.
This is done by tracking the largest row number that contains
non-empty pixels.
In unhook, when emitting the image, the image height is adjusted based
on this value.
This is to optimize sixel_add(). It already checks ‘pos’, to see if it
needs to resize the image, and if the resize fails (either due to
allocation failures, or because we’ve reached the maximum width), we
bail out.
We need to do the same thing for height. But, we don’t want to add
another check to sixel_add() since its’s performance critical.
By setting ‘col’ outside the image boundaries when ‘row’ reaches the
maximum image height, we can “re-use” the existing code in
sixel_add().
This reverts commit 5a9d70da837c20a9641d6cbadccc962a8e5a4283.
This broke jexer.
Comparing with what XTerm does, we can see that it updates its image
height for *each* pixel in *each* sixel. I.e. empty pixels at the
bottom of a sixel will not increase the image size.
Foot currently bumps the image height 6 pixels at a time, i.e. always
a whole pixel.
Given this, always rounding up the height to the nearest multiple of
6 (say, for example, when responding to a DECGRA command), is wrong.
Now, we use the image size specified in DECGRA as is, while still
allowing the image to grow beyond that if necessary.
What’s left to do, if we want to behave *exactly* like XTerm, is stop
growing the image with 6 pixels at a time when dynamically adjusting
the image size.
By storing the current row’s byte offset into the backing image in the
terminal struct.
This replaces the ‘imul’ with a load, which can potentially be
slow. But, this data should already be in the cache.
When we allocate the backing buffer, the number of allocated rows is a
multiple of 6.
When persisting the image, make sure its height does not exceed the
current maximum height.
But appears to have been allowed in older versions of meson.
Fixes an issue where foot’s version was always reported as the last
stable release, even though it was a git build.
Shift+Tab is an odd bird in XKB; it produces a shifted variant of tab,
ISO_Left_Tab, with Shift being a consumed modifier.
From a terminal perspective, Tab and ISO_Left_Tab are the same key,
with Shift+Tab generating ‘CSI Z’, and all other combos generating a
‘CSI 27;mods;9~’ escape.
Before, when we didn’t filter out consumed modifiers, we could simply
re-use the keymap for Tab when translating ISO_Left_Tab.
This is no longer possible, since shift has been filtered out. As a
result, Shift+Tab no longer generated ‘CSI Z’, but ‘\t’. I.e as if
shift wasn’t being pressed.
This patch adds a separate keymap table for ISO_Left_Tab. It’s MOD_ANY
entry corresponds to Shift+Tab, and emits ‘CSI Z’. All its other
entries *exclude* shift (since it has been consumed), *but*, the
encoded modifier (in the escape sequences) *include* shift.
This requirement is due to the recently added maybe_repair_key_combo()
function, which is making use of xkb_keymap_key_get_mods_for_level().
See also: commit a5b554761a.
Foot has, up until now, used a fixed image size when the application
used DECGRA (Raster Attributes) to “configure” the image size.
This was based on a misunderstanding, that this was how you emitted
sixels where the height was *not* a multiple of 6.
This isn’t the case. The VT340 documentation is actually pretty clear
about this:
Ph and Pv do not limit the size of the image defined by the sixel
data. However, Ph and Pv let you omit background sixel data from the
image definition and still have a color background. They also
provide a concise way for the application or terminal to encode the
size of an image.
This is also how XTerm behaves. Test image:
\EPq
"1;1;1;1
#0;2;0;0;0#1;2;100;100;0#2;2;0;100;0
#1~~@@vv@@~~@@~~$
#2??}}GG}}??}}??-
#1!14@
\E\
This uses DECGRA to set the image size to 1x1. The sixel however
is *not* clipped to 1x1, but is resized to 14x12
The ‘proto_chars’ variable is kind of like a FIFO, where characters
from the grid is pushed to the back of it. I.e. the last <n> bytes are
at the *end* of the array.
This was what the protocol matching logic expected. However, it was
only true after reading ‘max_prot_len’ characters from the
grid. Before that, the last <n> bytes from the grid was in
the *beginning* of ‘proto_chars’.
This meant we did not detect URLs starting in the top left corner of
the viewport. Test case:
foot sh -c "echo http://test.com && sleep 9999"
If a user-provided key combo contains both an explicit modifier, and
the shifted symbol it produces, replace the symbol with its un-shifted
version.
Example: Control+Shift+U contains both ‘Shift’, and ‘U’, where ‘Shift’
is the modifier that shifts ‘u’ to ‘U’.
Such modifiers are “consumed”, i.e. filtered out, when matching key
combos. As such, Control+Shift+U will never trigger since we’ll never
be able to match the consumed modifier ‘Shift’.
The above combo can be written in two ways:
- Control+U
- Control+Shift+u
To be able to detect that Control+Shift+U is an invalid combination,
we need to check all *shifted* symbols for all key *codes*.
Once we’ve detected a shifted symbol (and where it differs from the
un-shifted symbol), we loop all modifier sets that produce the shift
level where our shifted symbol is. For each such set, check if it
intersects with the user-provided key combo’s modifier set.
If there is an intersection, it means the user provided a key combo
containing a modifier and symbol, such that the modifier is consumed
when the symbol is produced.
In this case, we replace the symbol with its un-shifted version.
