This is equivalent to patches 7.4.396, 7.4.445 and 7.4.598.
vim-patch:7.4.396
Problem: When 'clipboard' is "unnamed", :g/pat/d is very slow.
(Praful)
Solution: Only set the clipboard after the last delete. (Christian
Brabandt)
1f285eb49a
vim-patch:7.4.445
Problem: Clipboard may be cleared on startup.
Solution: Set clip_did_set_selection to -1 during startup. (Christian
Brabandt)
1a19d37d90
vim-patch:7.4.598
Problem: ":tabdo windo echo 'hi'" causes "* register not to be
changed.
(Salman Halim)
Solution: Change how clip_did_set_selection is used and add
clipboard_needs_update and global_change_count. (Christian
Brabandt)
af6a579263
Co-Author: @bfredl
Problem: Commands like `:e term://top` display `term://.//PID:(null)`
instead of `term://.//PID:top`.
Solution: Remove the optional `name` argument (it allowed
ambiguous/potentially destructive cases, so it was problematic anyway).
Closes#3113Closes#2924
RETSIGTYPE, SIGRETURN, and SIGDEFARG are always defined the same, so
inline them for simplicity; SIGDUMMYARG, however, is unused, so remove
it.
Even if these things don't work on windows, the signal handler in
if_cscope.c is already wrapped with '#idef UNIX', as is its only call
site.
Reviewed-by: Justin M. Keyes <justinkz@gmail.com>
SIGHAS3ARGS appears in Vim's source code like this:
/* List 3 arg systems here. I guess __sgi, please test and correct me. jw. */
if defined(__sgi) && defined(HAVE_SIGCONTEXT)
define SIGHAS3ARGS
endif
Given we don't support SGI, remove it.
vim.h is too big, and these were missing parens around the arguments.
These may already be defined in <sys/param.h> or <windef.h>, so guard
them with an #ifndef.
extracted from #810
Problem: Accessing memory before an allocated block.
Solution: Check for not going before the start of a pattern. (Dominique Pelle)
https://github.com/vim/vim/commit/v7-4-799
In Windows there is no equivalent to the filesystem executable bit; the
documentation states that for Windows :executable() returns
1 for all files. But this behaviour was broken because is_executable()
checked for the UNIX bit.
When WIN32 is defined we now skip the S_IXUSR check.
After @250aca4f8938 it is possible that terminal_close will be called without
invoking the close_cb(which normally destroys the terminal structure). If this
happens, the terminal buffer will already be deleted so there's no need to call
`bwipeout!`.
The call to `event_init()` was too late. `command_line_scan()` in `main()` could
already need the loop initialized. Ref https://github.com/neovim/neovim/issues/3045#issuecomment-123405833.
A consequence of this change is that it was necessary to move the
`channel_from_stdio()` call to `command_line_scan()` when embedded_mode is set.
- Explicitly set the SignalWatcher event queue. Without this, the watcher will
publish events to the fast queue, resulting in resize bugs for certain
terminals(#2322).
- Set `async = false` to the `remote_ui_attach` handler(It was a deferred
before, this is the new equivalent)
It is possible for a processed event in the input loop of
terminal_enter() to destroy the terminal. But this is undetected by the
function and it still tries to use the freed terminal.
Use a reference count to delay the freeing of the terminal until
terminal_enter() returns.
Fixes#3112
os_file_is_readonly() in its current form is equivalent to
!os_file_is_writable(). This does not appear to be a bug, because Vim's
use of check_file_readonly() (which we changed to os_file_is_readonly())
is equivalent to !os_file_is_writable() in every case.
os_file_is_readonly() also fails this test:
returns false if the file is non-read, non-write
A more useful form would define behavior under these cases:
- path is executable (but not writable)
- path is non-existent
- path is directory
But there is no reason for os_file_is_readonly() to exist, so remove it.
Processing a stream's output can be queued. If stream_close() is called
before the queue is processed, the RBuffer containing the stream's data
is freed and the next read event would try to access freed memory.
To fix this behavior, use the stream's pending requests counter.
