| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
| |
An uv_pipe_t handle for a WStream could be left open for a
particular code path.
Patch by tarruda.
|
| |
|
|
|
| |
The argument argv of job_start() and channel_from_job() will be
freed. Mark them as such in the comments of this functions.
|
| |
|
|
|
| |
If a new job cannot be started because no slots are free, we return early
without freeing the argv argument.
|
| |
|
|
|
|
| |
- When an invalid msgpack RPC msg is received from a channel
we now close that channel all calls on that channel fail with
an error message.
|
| | |
|
| | |
|
| |
|
|
|
| |
- When validating a msgpack msg we need to return on the first error
otherwise we can SEGFAULT with invalid checks
|
| |
|
|
|
|
|
|
| |
Use save_tv_as_string(), same as vimL system(). This also makes
jobsend() more liberal in what it can accept. For example,
`jobsend(j, 123)` is now valid.
Closes #1176
|
| |
|
|
|
|
|
|
| |
Factor out string_to_list() from f_system()'s implementation
and use that to set job_data. This has the technical advantage of
preserving NULs, and may be more convenient for users.
Required for #1176.
|
| |
|
|
| |
Only read up to the last newline in push_job_event().
|
| |
|
|
|
| |
Needed for accessing this information without direct access to the
rstream_buffer.
|
| | |
|
| |
|
|
|
| |
stdout/stderr should only be closed after the job truly exits, or else we can
lose data sent by it.
|
| | |
|
| |
|
|
|
| |
Replace long with int64_t.
Apart from being convenient, not using long is required by clint.
|
| |
|
|
| |
Replace int with bool where appropriate.
|
| |
|
|
|
| |
This, in addition to being more correct, allows removing a lot of
explicit casts.
|
| |
|
|
| |
Unused since 457bb2615154946d273d75e07f5d5a936f50ede0.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- memfile_defs.h:
* hashtab_T: mht_mask: long_u -> size_t.
Masks are used to truncate keys to a value adequate for an index
in the array of entries. Value of the mask plus one is the
current size of the array. Both of those reasons imply the
soundness of size_t for this type.
* hashtab_T: mht_count: long_u -> size_t.
- memfile.c:
* total_mem_used: long_u -> size_t.
* mf_hash_free_all: idx: long_u -> size_t.
* mf_hash_add_item: idx: long_u -> size_t.
* mf_hash_find: idx: long_u -> size_t.
* mf_hash_grow: i: long_u -> size_t.
* mf_hash_grow: j: long_u -> size_t.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
- Add memfile.c to converted files list.
- Fix conversion issues:
* bhdr_T : bh_page_count : int -> unsigned.
* bhdr_T : bh_flags : char -> unsigned.
* mf_new() : page_count : int -> unsigned.
* mf_get() : page_count : int -> unsigned.
* mf_release() : page_count : int -> unsigned.
* mf_alloc_bhdr() : page_count : int -> unsigned.
* mf_trans_add() : page_count : int -> unsigned.
* mf_put() : flags : int -> unsigned.
|
| |
|
|
|
|
| |
- Drop '_S' suffix for struct names.
- Make struct names match corresponding type's name (just removing '_S' suffix).
- Rename NR_TRANS type/struct (just ugly).
|
| |
|
|
|
|
|
|
|
|
|
|
| |
memfile_defs.h:
- Inline struct definitions in typedefs.
- Move memfile_T definition to this file (weirdly, was in buffer_defs.h).
memfile.c:
- Use C99 style variable declarations. This is, move variable declarations as
near to first-usage point as possible).
- Modernize old-style function declarations.
- Fix indent at some places (some multiline expressions and the like).
|
| |
|
|
|
| |
- Restyle comments (/// when appropiate, // otherwise).
- Improve comments (add new comments, augment/clarify existing ones).
|
| |
|
|
|
| |
Note: No effort is done on refactoring anything in eval module, as it
will be eventually dropped.
|
| |\
| |
| | |
version: Refer to nvim help on startup screen.
|
| | | |
|
| |\ \
| |/
|/| |
re-enable iconv
|
| | |
| |
| |
| |
| | |
This seems to have been disabled in the transition from vim to neovim,
re-enable it.
|
| |/
|
|
|
|
|
|
|
|
| |
It is currently possible for a client to send a response that doesn't match the
current server->client request(at the top of the stack). This commit fixes that
by delaying notifications to until the first `channel_send_call` invocation
returns.
Also remove the "call stack" size check, vim will already break if the call
stack goes too deep.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
