Problem: win32: issues with stable python ABI
Solution: if_python3,win32: Fix Python3 stable ABI
There were some issues in current stable ABI implementation on Windows:
* Python DLL name should be `python3.dll` instead of `python311.dll` and
so on. (See: https://docs.python.org/3/c-api/stable.html)
* Some non-stable API functions were used:
- `_PyObject_NextNotImplemented`
- `PyStdPrinter_Type`
* `reset_stdin()` and `hook_py_exit()` didn't work with `python3.dll`.
`python3.dll` is a special type of DLL called forwarder DLL.
It just forwards the functions to other DLL (e.g. `python311.dll`).
There were two issues regarding these functions:
- `python3.dll` doesn't have import tables. This caused a crash in
`get_imported_func_info()`. Add a check whether the specified DLL
has an import table.
- `reset_stdin()` and `hook_py_exit()` should be applied to the
forwarded DLL (e.g. `python311.dll`), not to `python3.dll`.
Check the export directory of `python3.dll` to find the forwarded
DLL and apply the functions to it.
closes: #13260
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Ken Takata <kentkt@csc.jp>
Problem: Cirrus CI disabled
Solution: re-enable Cirrus CI
Ref patch 9.0.1912:
> Perhaps at the beginning of the next month we can revisit and enable
> just a build without testing it. Hopefully this is won't take too
> many credits and we can at least verify that building works.
Actually enabling testing should be fine. In the last month there were
three Cirrus CI jobs and credits ran out on Sep 15, but now there is
only one Cirrus CI job, so credits shouldn't run out.
closes: #13261
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: No filetype detection for just files
Solution: Detect just files (*.just, justfile, etc)
closes: #13271
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: dundargoc <gocdundar@gmail.com>
Problem: Vim9: object members can change type
Solution: Check type during assignment to object/class var
closes: #13127closes: #13262
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: xattr: permission-denied errors on write
Solution: ignore those errors
closes: #13246
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Gene C <arch@sapience.com>
* Updated to Vim 9.0.1968. Several refinements and corrections
* typo fixed
Co-authored-by: RestorerZ <restorer@mail2k.ru>
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Clean up cmdline option completion code
Solution: Fix various minor problems
- Fix manual array size calculations to just use `ARRAY_LENGTH()`.
- Fix unintentional typo in comments due to copy-paste error.
- Fix assert_equal() usages to pass the expected value to first
parameter instead of 2nd one to avoid confusion.
- Fix signed vs unsigned warnings
- Correct misplaced comments about set_op_T and set_prefix_T
and fix a typo in another comment
closes: #13249closes: #13237
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: win32: missing '**' expansion test (after v9.0.1947)
Solution: Add test for MS-Windows
win32: Add "**" test
Vim supports "**" on MS-Windows. However, it is not tested by
`Test_glob_extended_bash`.
Unlike Unix, it doesn't use 'shell' and doesn't support {,} expansion.
So, I added as a separate test.
related: #13205closes: #13250
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Ken Takata <kentkt@csc.jp>
Problem: macOS: FEAT_SOUND guard too restrictive
Solution: check for older macOS support properly
Fix macOS FEAT_SOUND guards to be less restrictive
This allows +sound to work on older macOS platforms again. The +sound
implementation uses APIs available in 10.6, but the code itself uses
generics with type parameters which was only added in Xcode 7 / clang 7,
which was released for macOS 10.11. This means as long as Vim is
compiled under 10.11+, and using a deployment target >= 10.6, the
feature will work.
closes: #13251
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
Problem: win32: high-dpi support can be improved
Solution: implement WM_GETDPISCALEDSIZE
win32: Better support of Per-Monitor V2 High DPI
Implement WM_GETDPISCALEDSIZE.
It is not so easy to calculate the new size without actually changing
the size. So, this returns an approximate size.
This doesn't work well when a bitmap font (e.g. FixedSys) is selected,
but I think this is acceptable.
closes: #11812closes: #13252
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Ken Takata <kentkt@csc.jp>
Problem: buffer-overflow in trunc_string()
Solution: Add NULL at end of buffer
Currently trunc_string() assumes that when the string is too long,
buf[e-1] will always be writeable. But that assumption may not always be
true. The condition currently looks like this
else if (e + 3 < buflen)
[...]
else
{
// can't fit in the "...", just truncate it
buf[e - 1] = NUL;
}
but this means, we may run into the last else clause with e still being
larger than buflen. So a buffer overflow occurs.
So instead of using `buf[e - 1]`, let's just always
truncate at `buf[buflen - 1]` which should always be writable.
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: cmdline completion should consider key option
Solution: Disable cmdline completion for key option, slightly
refactor how P_NO_CMD_EXPAND is handled
Harden crypto 'key' option: turn off cmdline completion, disable set-=
"set-=" can be used maliciously with a crypto key, as it allows an
attacker (who either has access to the computer or a plugin author) to
guess a substring by observing the modified state. Simply turn off
set+=/-=/^= for this option as there is no good reason for them to be
used.
Update docs to make that clear as well.
Also, don't allow cmdline completion for 'key' as it just shows *****
which is not useful and confusing to the user what it means (if the user
accidentally hits enter they will have replaced their key with "*****"
instead).
Move logic to better location, don't use above 32-bit for flags
Move P_NO_CMD_EXPAND to use the unused 0x20 instead of going above
32-bits, as currently the flags parameter is only 32-bits on some
systems. Left a comment to warn that future additions will need to
change how the flags work either by making it 64-bit or split into two
member vars.
Also, move the logic for detecting P_NO_CMD_EXPAND earlier so it's not
up to each handler to decide, and you won't see the temporary "..." that
Vim shows while waiting for completion handler to complete.
closes: #13224
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
Problem: xattr errors not translated
Solution: mark for translation, consistently capitalize
first letter.
closes: #13236
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: configure prints stray 6 when checking libruby
Solution: redirect stdout to dev/null
configure: Do not print "6" when checking for libruby
`expr` will print the matched string length to the standard output.
Current `configure` output looks like this:
```
checking Ruby header files... /usr/include/ruby-3.1.0
6
```
The script really only cares about `expr` exit code.
closes: #13234
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Illia Bobyr <illia.bobyr@gmail.com>
Problem: wrong auto/configure script
Solution: regenerate with autoconf 2.71
configure: sys/xattr.hs: Regenerate with autoconf 2.71
It seems that `auto/configure` update in
commit 6de4e58cf27a3bb6e81653ca63b77e29d1bb46f2 (tag: v9.0.1963)
Author: zeertzjq <zeertzjq@outlook.com>
Date: Sat Sep 30 14:19:14 2023 +0200
patch 9.0.1963: Configure script may not detect xattr
Problem: Configure script may not detect xattr correctly
Solution: include sys/xattr instead of attr/xattr,
make Test_write_with_xattr_support() test
xattr feature correctly
This also applies to the Smack security feature, so change the include
and configure script for it as well.
closes: #13229
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
was done manually, and missed an update to the generated variable name.
closes: #13235
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Illia Bobyr <illia.bobyr@gmail.com>
Problem: xattr support fails to build on MacOS X
Solution: Disable xattr support for MacOS X
MacOS X uses the same headers and functions sys/xattr.h but the function
signatures for xattr support are much different, so building fails.
So let's for now disable xattr support there.
closes: #13230closes: #13232
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Configure script may not detect xattr correctly
Solution: include sys/xattr instead of attr/xattr,
make Test_write_with_xattr_support() test
xattr feature correctly
This also applies to the Smack security feature, so change the include
and configure script for it as well.
closes: #13229
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: No support for writing extended attributes
Solution: Add extended attribute support for linux
It's been a long standing issue, that if you write a file with extended
attributes and backupcopy is set to no, the file will loose the extended
attributes.
So this patch adds support for retrieving the extended attributes and
copying it to the new file. It currently only works on linux, mainly
because I don't know the different APIs for other systems (BSD, MacOSX and
Solaris). On linux, this should be supported since Kernel 2.4 or
something, so this should be pretty safe to use now.
Enable the extended attribute support with normal builds.
I also added it explicitly to the :version output as well as make it
able to check using `:echo has("xattr")`, to have users easily check
that this is available.
In contrast to the similar support for SELINUX and SMACK support (which
also internally uses extended attributes), I have made this a FEAT_XATTR
define, instead of the similar HAVE_XATTR.
Add a test and change CI to include relevant packages so that CI can
test that extended attributes are correctly written.
closes: #306closes: #13203
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Cmdline completion for 'listchars' fields doesn't include
"multispace" and "leadmultispace" (after 9.0.1958).
Solution: Include "multispace" and "leadmultispace" in lcstab.
closes: #13225
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: Make CI checks more strict
Solution: Add -Wstrict-prototypes -Wmissing-prototypes to CI,
fix uncovered problems
Add -Wstrict-prototypes -Wmissing-prototypes warnings check to CI
Add two new warnings to CI, silence some Perl related build-warnings:
- `strict-prototypes` helps prevent declaring a function with an empty
argument list, e.g. `int func()`. In C++, that's equivalent to `int
func(void)`, but in C, that means a function that can take any number
of arguments which is rarely what we want.
- `missing-prototypes` makes sure we use `static` for file-only internal
functions. Non-static functions should have been declared on a
prototype file.
- Add `no-compound-token-split-by-macro` to the perl cflags, since it
throws out a bunch of perl-related warnings that make the CI log
unnecessary verbose and hard to read. This seems to happen only with
clang 12 and above.
When applying those changes, it already uncovered a few warnings, so fix
up the code as well (fix prototypes, make the code static, remove
shadowed var declaration)
GTK header needs to have #pragma warning suppressiong because GTK2
headers will warn on `-Wstrict-prototypes`, and it's included by gui.h
and so we can't just turn off the warning in a couple files.
closes: #13223closes: #13226
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
Problem: Vim9: methods parameters and types are covariant
Solution: Support contra-variant type check for object method arguments
(similar to Dart).
closes: #12965closes: #13221
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: cannot complete option values
Solution: Add completion functions for several options
Add cmdline tab-completion for setting string options
Add tab-completion for setting string options on the cmdline using
`:set=` (along with `:set+=` and `:set-=`).
The existing tab completion for setting options currently only works
when nothing is typed yet, and it only fills in with the existing value,
e.g. when the user does `:set diffopt=<Tab>` it will be completed to
`set diffopt=internal,filler,closeoff` and nothing else. This isn't too
useful as a user usually wants auto-complete to suggest all the possible
values, such as 'iblank', or 'algorithm:patience'.
For set= and set+=, this adds a new optional callback function for each
option that can be invoked when doing completion. This allows for each
option to have control over how completion works. For example, in
'diffopt', it will suggest the default enumeration, but if `algorithm:`
is selected, it will further suggest different algorithm types like
'meyers' and 'patience'. When using set=, the existing option value will
be filled in as the first choice to preserve the existing behavior. When
using set+= this won't happen as it doesn't make sense.
For flag list options (e.g. 'mouse' and 'guioptions'), completion will
take into account existing typed values (and in the case of set+=, the
existing option value) to make sure it doesn't suggest duplicates.
For set-=, there is a new `ExpandSettingSubtract` function which will
handle flag list and comma-separated options smartly, by only suggesting
values that currently exist in the option.
Note that Vim has some existing code that adds special handling for
'filetype', 'syntax', and misc dir options like 'backupdir'. This change
preserves them as they already work, instead of converting to the new
callback API for each option.
closes: #13182
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yee Cheng Chin <ychin.git@gmail.com>
Problem: termcap options should change on keyprotocol setting
Solution: Apply termcap entries when 'keyprotocol' changes
When the 'keyprotocol' option was set after startup (including in a
user's .vimrc) the termcap entries associated with the matching protocol
were not applied. Thus, setting the option has no affect.
When 'keyprotocol' is changed it should also update the termcap entries.
closes: #13211
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Gregory Anders <greg@gpanders.com>
Problem: Custom cmdline completion skips original cmdline when pressing
Ctrl-P at first match if completion function invokes glob().
Solution: Move orig_save into struct expand_T.
closes: #13216
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: Vim9: lockvar issues with objects/classes
Solution: fix `get_lhs()` object/class access and avoid `SEGV`,
make error messages more accurate.
- `get_lval()` detects/returns object/class access
- `compile_lock_unlock()` generate code for bare static and obj_arg access
- `do_lock_var()` check lval for `ll_object`/`ll_class` and fail if so.
Details:
- Add `ll_object`/`ll_class`/`ll_oi` to `lval_T`.
- Add `lockunlock_T` to `isn_T` for `is_arg` to specify handling of `lval_root` in `get_lval()`.
- In `get_lval()`, fill in `ll_object`/`ll_class`/`ll_oi` as needed; when no `[idx] or .key`, check lval_root on the way out.
- In `do_lock_var()` check for `ll_object`/`ll_class`; also bullet proof ll_dict case
and give `Dictionay required` if problem. (not needed to avoid lockvar crash anymore)
- In `compile_lock_unlock()` compile for the class variable and func arg cases.
closes: #13174
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Ernie Rael <errael@raelity.com>
Problem: CI: change netrw label in labeller bot
Solution: Rename it to 'plugin-netrw'
closes: #13217
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Ken Takata <kentkt@csc.jp>
Problem: Misplaced comment in errors.h
Solution: Move it up
closes: #13218
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Ken Takata <kentkt@csc.jp>
Problem: Vim9: hard to debug vim9_class errors from CI
Solution: Include the line number in assert_xxx() calls. Include the
entire error message in the tests. Fix the indentation in the
test file. Add tags for new error codes.
closes: #13206
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: Vim9: error codes spread out
Solution: group them together and reserve 100
more for future use
Reserve 100 error codes for future enhancements to the Vim9 class
support
closes: #13207
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: Vim9: object variable "this." should only be used in
constructor
Solution: Disallow to this in normal object methods (other than
constructors)
closes: #13152closes: #13212
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: h-east <h.east.727@gmail.com>
Problem: Bash Expansion test fails on Windows/MacOS
Solution: Disable Test_glob_extended_bash for now
This test doesn't work on Windows even if bash can be executed, since
the globstar functionality has only been enabled in Unix builds of Vim
(Commit 9eb1ce531527a7177d16373b0f8689bbcd3d5f73, patch 9.0.1946).
closes: #13205
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Ken Takata <kentkt@csc.jp>
Problem: filename expansion using ** in bash may fail
Solution: Try to enable the globstar setting
Starting with bash 4.0 it supports extended globbing using the globstar
shell option. This makes matching recursively below a certain directory
using the ** pattern work as expected nowadays. However, we need to
explicitly enable this using the 'shopt -s globstar' bash command.
So let's check the bash environment variable $BASH_VERSINFO (which is
supported since bash 3.0 and conditionally enable the globstar option,
if the major version is at least 4. For older bashs, this at least
shouldn't cause errors (unless one is using really ancient bash 2.X or
something).
closes: #13002closes: #13144
Signed-off-by: Christian Brabandt <cb@256bit.org>
Problem: Vim9: missing support for ro-vars in interface
Solution: Support only read-only object variables in an interface,
add additional checks when parsing class definitions.
closes: #13183
cloess: #13184
cloess: #13185.
closes: #13188
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: Vim9: function instruction pointer invalidated
Solution: Use the funcref index instead of the instruction pointer
closes: #13178closes: #13196
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: CI not run with clang-17
Solution: Update CI to use clang-17
closes: #12745
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Philip H <47042125+pheiduck@users.noreply.github.com>
Problem: Vim9: execution stack invalidated with null object
Solution: Check for a null object before adjusting the execution stack
closes: #13186
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: still a problem when processing LSP RPC requests
Solution: When processing async LSP RPC requests, compare sequence
numbers only in response messages
A LSP request message can be sent to the language server either
synchronously (ch_evalexpr) or asynchronously (ch_sendexpr). In both
cases, when looking for response messages by using the sequence number,
LSP requests messages from the language server with the same sequence
number should not be used. Patch 9.0.1927 fixed this issue for
synchronous requests. This PR fixes the issue for asynchronous requests
and adds additional tests.
closes: #13158
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Yegappan Lakshmanan <yegappan@yahoo.com>
Problem: multispace wrong when scrolling horizontally
Solution: Update position in "multispace" or "leadmultispace" also in
skipped chars. Reorder conditions to be more consistent.
closes: #13145closes: #13147
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: missing test for mouse click + 'virtedit'
Solution: Add test for clicking after eol with 'virtualedit' and wrapped
line
closes: #13157
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: test: using wrong expected message in test_crypt
Solution: make use of single quotes
closes: #13151
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: Isao Sato <svardew@gmail.com>
Problem: Vim9: not consistent error messages
Solution: Make error messages more consistent. Use "variable" for
(object/class) member
closes: #13155
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: RestorerZ <restorer@mail2k.ru>
Problem: :bwipe fails after switching window from aucmd_win.
Solution: Decrement b_nwindows after switching back to aucmd_win.
closes: #13160
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
Problem: Can change the type of a v: variable using if_lua.
Solution: Add additional handling of v: variables like :let.
closes: #13161
Signed-off-by: Christian Brabandt <cb@256bit.org>
Co-authored-by: zeertzjq <zeertzjq@outlook.com>
