| Age | Commit message (Collapse) | Author |
|---|
|
suggested in [0] since it will be help for latex
[0]: https://github.com/martanne/vis/commit/dac6a7e#comments
|
|
|
|
|
|
|
|
{w,}ctype(3) character classes are essentially broken for non-ascii
text. 711447a tried to fix this for words surrounded by blanks but forgot
the use case of completing function and variable names in source code.
Instead of relying on the terrible ctype interface we can hand pick
a set that is good enough for both source code completion and writing
prose. This set is consistent with the old [:alnum:] behaviour for ascii
text but also supports words with single width non-ascii characters.
fixes 711447a: vis-complete: handle non-ascii text
closes #1132: Source code completion are broken
|
|
The temporary directory for vis-single was hard coded to /tmp.
If /tmp happens to be mounted noexec then vis fails as it cannot
run anything placed inside the temporary directory. If the TMPDIR
environment variable is set, respect it for vis-single.
|
|
|
|
|
|
|
|
|
|
fixes #1119: lua: lpeg module isn't actually optional
|
|
`SIZE_MAX` cannot be represented accurately in `lua_Number`. A correct
solution probably doesn't exist but we can silence the warning by
explicitly casting to `lua_Number` and changing the comparison to `<`
instead of `<=`.
checkpos() deals with large numbers for file ranges. For most users we
can assume no one is editing files that are `SIZE_MAX` bytes long (many
petabytes). For obscure systems where `SIZE_MAX` is a small number this
will result in a maximum range (in lua) of 1 byte less than before.
fixes #1120: vis-lua.c:504:21: warning: implicit conversion changes value
|
|
This is in response to a comment left on a35e7ea. Backwards compatibility
is a good idea for at least a release.
|
|
|
|
I thought I fixed these in the applied patch but I guess they slipped by
|
|
Rationale
A modern text editor usually includes tools for helping user
to avoid mistakes in texts. Those tools include spell checkers and
programming language integrations. Though vis explicitly states
that the full featured IDE is not a goal, implementing some of
the tools might be achieved using its Lua API. Unfortunatelly
the API misses the ability to start a process and to perform
a communication with it without completely blocking the editor UI,
which is crucial for any tool that performs background tracking of
the inserted text (e. g. language servers).
Implementation details
New feature introduces new API method: communicate. The method
start a new process and returns a handle to communicate with
the process instantly. The patch inserts stderr and stdout
file descriptors of the process to the pselect call of the main loop
used for reading user input to track the process state without
blocking the main loop until the process is finished.
Any changes in the process state cause the iteration of the main loop
and are being exposed to the Lua API as new event: PROCESS_RESPONSE.
|
|
|
|
The first point of this commit is to allow all options to be read from
lua. This has a number of uses for plugin writers. They are grouped into
a couple of tables depending on what they control:
`vis.options`: table with global configuration
`win.options`: table with window specific configuration
The second point is to allow you to set all these options as if they
were simply lua variables. Technically this is already possible by
using `vis:command("set ...")` but personally I think this interface
is cleaner. Note that this already possible for some things like the
current mode (eg. vis.mode = vis.modes.VISUAL). Examples:
`vis.options.ai = true`
`win.options.brk = " !?."`
`win.options = { showeof = true, showtabs = true }
There are a number of related issues and pull requests:
closes #803: Lua API: let plugins read the values of options
closes #812: Window layout property
supersedes/closes #717: Add ability to access tabwidth from Lua
supersedes/closes #1066: expose UI layout and allow it to be set from lua API
|
|
|
|
some users were (rightfully) annoyed by this
|
|
from feature_test_macros(7):
> Defining _XOPEN_SOURCE with a value of 700 or greater produces the
> same effects as defining _POSIX_C_SOURCE with a value of 200809L or
> greater.
Depending on the configuration and system pkg-conf files there can be
redefinition warnings. Rather than patching with a -U_POSIX_C_SOURCE
it can just be dropped instead.
|
|
|
|
The '[:alnum:]' set does not include non-ascii text which results in
the non-ascii text being replaced with newlines. Using the '[:blank:]'
set with no complement flag fixes this issue.
|
|
Before we were not taking non-ascii characters into account properly. With
this patch we still mix byte counts and "grapheme cluster" (i.e. complete
glyphs that are rendered in a terminal cell) counts but the code should
be less broken in the more common case now.
|
|
|
|
The only place where this behaviour was encountered was in
file_lines_iterator() and it was just being worked around.
|
|
Reading from curs_refresh(3X) from curses, calling doupdate() repeatedly
will cause 'several bursts of output to the screen'. wnoutrefresh() has
the smarts to only copy the changed lines to the copied virtual screen,
but doupdate() does not.
There have been several bug reports related to flickering but all seems
to be inconsistenly reproducible due to different terminal buffering
behavior. See #1032, #327
Unfortunately, when I am using a slow display, I still notice
flickering, so this commit changes the routines for opening new windows
and splitting windows to wait until the last change is finished before
calling doupdate().
|
|
It might be more comfortable to navigate through the files
arranged vertically.
Add ability to specify VIS_OPEN_LINES environment variable
which is passed to vis-menu as -l option.
It would be better to add vis option for this to set it
via lua config, but it might be added later.
Signed-off-by: Vadym Kochan <vadim4j@gmail.com>
|
|
this is contolled by the wrapcolumn/wc and breakat/brk options
related #142: Word wrap and line breaks
related #932: Vis for Prose?
related #1092: Disabling line wrapping
|
|
|
|
Aside from the possibility of future syntax highlighting (similar to
https://github.com/vim/vim/blob/master/runtime/syntax/mail.vim in Vim).
After objections on the list, I have decided not to include the conversion
and user has to introduce it in their visrc.lua:
vis.ftdetect.filetypes.mail.cmd = { "x/\r/ d" }
|
|
When `file` was made optional the variable was changed to be initialized
by `vis->win->file` instead of a known safe file pointer. If `vis:pipe()`
is called based on the event `FILE_OPEN` the `file` parameter can be
valid even when `vis->win` is not yet present. Assuming `file` was
provided this would be handled later on if vis didn't segfault. By
initializing to NULL when `vis->win` isn't present `file` is given
a chance be handled later. In the case where `file` wasn't given and
`vis->win` is missing an error is thrown and `vis:pipe()` exits.
fixes #1107 - Lua API: vis:pipe() causes a segfault when called before
a window is open
|
|
This enables restoring the terminal from a fullscreen command like
curses based program. Use cases are e.g. a file picker based on some
external program like nnn (https://github.com/jarun/nnn).
|
|
this slipped through in commit 6be370d
|
|
Fixes #1060 - :help doesn't display mappings starting with <Space>
correctly
Co-authored-by: Randy Palamar <palamar@ualberta.ca>
|
|
text_paragraph_prev():
Bring back the previous usage of text_iterator_byte_get() in the
while conditional and text_iterator_char_prev() in the loop body.
Fixes #1028 - { moves back a paragraph too much if cursor at start of line
|
|
|
|
fixes #971
|
|
We can make `od` skip the address radix, so `sed` does not
need to remove it.
|
|
|
|
this seems to be broken if the actor isn't the owner of the repository
|
|
this is mostly useful for the internal vis usage and makes both `*` and
`+` registers work on macOS/cygwin.
fixes: #1067
|
|
|
|
(horizontally) and (vertically) were kept to be consistent with the 'v'
mnemonic and with the enum labels in the code.
|
|
this is more stable than grabbing from the CVS web interface
|
|
|
|
|
|
the message window gets closed like a normal window and nothing calls
this function. In fact, it wasn't even used when it was added 8 years
ago in 979ab79.
|
|
When the cmd closes the window the window pointer gets freed along the
way. We can't use win->next to update the loop variable if sam_execute()
has been called. Instead we can store win->next early and use that
variable to continue the loop.
fixes #1090
|
|
|
