Work around a glibc bug where it truncates the argument to fpclassify()
from double to float by implementing fpclassify() ourselves.
Correctness test (Note that the FP_SUBNORMAL test depends on an atof() that
knows how to parse subnormals. Glibc does, not sure about other libcs.):
#include <math.h>
#include <stdint.h>
#include <string.h>
int xfpclassify(double d)
{
uint64_t m;
int e;
memcpy(&m, &d, sizeof(m));
e = 0x7ff & (m >> 52);
m = 0xfffffffffffffULL & m;
switch (e) {
default: return FP_NORMAL;
case 0x000: return m ? FP_SUBNORMAL : FP_ZERO;
case 0x7ff: return m ? FP_NAN : FP_INFINITE;
}
}
#include <assert.h>
#include <stdlib.h>
int main(void)
{
assert(FP_ZERO == xfpclassify(atof("0.0")));
assert(FP_ZERO == xfpclassify(atof("-0.0")));
assert(FP_NORMAL == xfpclassify(atof("1.0")));
assert(FP_NORMAL == xfpclassify(atof("-1.0")));
assert(FP_INFINITE == xfpclassify(atof("inf")));
assert(FP_INFINITE == xfpclassify(atof("-inf")));
assert(FP_NAN == xfpclassify(atof("nan")));
assert(FP_NAN == xfpclassify(atof("-nan")));
assert(FP_SUBNORMAL == xfpclassify(atof("1.8011670033376514e-308")));
return 0;
}
Problem: Escaping for fish shell is skipping some characters.
Solution: Escape character after backslash if needed. (Jason Cox,
closesvim/vim#8827)
6631597452
I mistakenly suggested maxlines=&cmdwinheight, forgetting that it is
calculated from topline, not cursor. maxlines=1 makes the most sense in
cmdwin.
ref #15401622a36b1f1
Add a new default autocommand to limit syntax highlighting
synchronization in the command window. This refactors the nvim_terminal
autocommand out of main() and into a new init_default_autocmds()
function, which is now part of the startup process and can be further
extended with more default autocommands down the road.
ref #6289#6399
Problem:
jobwait() returns early if the job was stopped, but the job might have
pending callbacks on its event queue which are required to complete its
teardown. State such as term->closed might not be updated yet (by the
pending callbacks), so codepaths such as :bdelete think the job is still
running.
Solution:
Always flush the job's event queue before returning from jobwait().
ref #15349
fixes#15524
Note: this is obviously a quickfix. A scalabe solution will
involve being able to specify a _list_ of modules to be
put into packages.preload, without needing to manually copypasta
a blurb of C code. Perhaps even involving bytecode for
static builds (to speedup initialization)
* fix(tutor): adjust over-80ch lines and corresponding expect file
* fix(tutor): standardise indentation and formatting, add nowrap modeline
- unifies the formatting/layout, which was a bit inconsistent,
- adds a nowrap modeline
Since the tutor uses a lot of conceals, which are included in the character
count when calculating line wrapping, lines were breaking at what looked like
odd spots, which gives a poor first impression and lowered readability.
I have adjusted some lines to be over 80ch in the source, but once they're
rendered out with conceals, they're actually under 80, so even with nowrap we
don't visually extend past 80.
fix#15088
refactor(codebase): Update uncrustify config and apply it on one file screen.c
The plan is to go through the codebase in batches, and remove the most egregious "make lint"
backlog failures, by applying this config.
Converter functions use a heap-allocated stack to handle complex
nested objects. However, these are often called with simple,
primitive values like integers or bools wrapped in an Object.
Avoid the memory allocation in this case using kvec_withinit_t
These functions do not involve msgpack. Initially the nvim api was
sometimes called the "msgpack API", but entry points from vim script
and lua are equally valid (and don't need to reference "msgpack")
After a process's refcnt is decremented to zero, it enqueues a
`process_close_event` on its own event queue. In `process_wait`, this
event should be processed immediately so that any process close
callbacks are executed before `process_wait` returns.
Update `process_wait` to always process the process's event queue after
the process is freed, rather than the event queue passed in as an
argument.
Remove the trailing slashes from 'undofile' and 'backupdir' before
creating directories. They cause problems on Windows which doesn't
recognize these slashes as proper path separators.
Copy the behavior of 'undodir' and create the last specified directory
in the 'backupdir' option if it doesn't exist.
Use trailing slashes for 'backupdir' as well as 'viewdir' and 'undodir'
by default. Note that 'undodir' always behaves as though it has the
trailing slashes, regardless of whether or not they are present. They
are added to the default option value to minimize surprise.
The '.' value in 'backupdir' is kept because the default behavior for
backups is solely to have a backup if the save of the main file to disk
fails. As soon as that save is completed the backup file is removed, so
generally there is no need to put them in a central location.
Co-authored by: murphy66 <murphy66@gmail.com>
Resolve an issue with deferred clearing of highlight failing if the
buffer is deleted before the timeout by checking whether the
buffer is valid first.
-range=-1 requires the current file to have at least <count> lines, whereas -addr=other doesn't.
-addr=other also sets <count> to -1 by default when it is not specified, though this feature seems undocumented.
Problem:
"set filetype=man" assumes the user wants :Man features, this does extra
stuff like renaming the buffer as "man://".
Solution:
- old entrypoint was ":set filetype=man", but this is too presumptuous #15487
- make the entrypoints more explicit:
1. when the ":Man" command is run
2. when a "man://" buffer is opened
- remove the tricky b:man_sect checks in ftplugin/man.vim and syntax/man.vim
- MANPAGER is supported via ":Man!", as documented.
fixes#15487
