Merge 'Refactor WStream to enable writing the same buffer to multiple targets'

scripts/msgpack-gen.lua

@@ -62,6 +62,12 @@ for i = 1, #arg - 1 do
     api.functions[#api.functions + 1] = tmp[i]
     local fn_id = #api.functions
     local fn = api.functions[fn_id]
+    if #fn.parameters ~= 0 and fn.parameters[1][2] == 'channel_id' then
+      -- this function should receive the channel id
+      fn.receives_channel_id = true
+      -- remove the parameter since it won't be passed by the api client
+      table.remove(fn.parameters, 1)
+    end
     if #fn.parameters ~= 0 and fn.parameters[#fn.parameters][1] == 'error' then
       -- function can fail if the last parameter type is 'Error'
       fn.can_fail = true
@@ -111,7 +117,7 @@ end
 output:write([[
 };
 
-void msgpack_rpc_dispatch(uint64_t id, msgpack_object *req, msgpack_packer *res)
+void msgpack_rpc_dispatch(uint64_t channel_id, msgpack_object *req, msgpack_packer *res)
 {
   Error error = { .set = false };
   uint64_t method_id = (uint32_t)req->via.array.ptr[2].via.u64;
@@ -121,7 +127,7 @@ void msgpack_rpc_dispatch(uint64_t id, msgpack_object *req, msgpack_packer *res)
       msgpack_pack_nil(res);
       // The result is the [channel_id, metadata] array
       msgpack_pack_array(res, 2);
-      msgpack_pack_uint64(res, id);
+      msgpack_pack_uint64(res, channel_id);
       msgpack_pack_raw(res, sizeof(msgpack_metadata));
       msgpack_pack_raw_body(res, msgpack_metadata, sizeof(msgpack_metadata));
       return;
@@ -169,8 +175,19 @@ for i = 1, #api.functions do
     output:write(fn.return_type..' rv = ')
   end
 
-  -- write the call without the closing parenthesis
-  output:write(fn.name..'('..call_args)
+  -- write the function name and the opening parenthesis
+  output:write(fn.name..'(')
+
+  if fn.receives_channel_id then
+    -- if the function receives the channel id, pass it as first argument
+    if #args > 0 then
+      output:write('channel_id, '..call_args)
+    else
+      output:write('channel_id)')
+    end
+  else
+    output:write(call_args)
+  end
 
   if fn.can_fail then
     -- if the function can fail, also pass a pointer to the local error object

scripts/run-api-tests.exp

@@ -28,9 +28,9 @@ send {
   for group in split(groups)
     exe 'augroup '.group
     autocmd!
-    augroup NONE
-    exe 'augroup! '.group
+    augroup END
   endfor
+  autocmd!
   tabnew
   let curbufnum = eval(bufnr('%'))
   redir => buflist

src/nvim/api/buffer.c

@@ -121,6 +121,7 @@ end:
     }
 
     free(rv.items);
+    rv.items = NULL;
   }
 
   return rv;

src/nvim/api/vim.c

@@ -7,6 +7,7 @@
 #include "nvim/api/private/helpers.h"
 #include "nvim/api/private/defs.h"
 #include "nvim/api/buffer.h"
+#include "nvim/os/channel.h"
 #include "nvim/vim.h"
 #include "nvim/buffer.h"
 #include "nvim/window.h"
@@ -327,6 +328,24 @@ void vim_set_current_tabpage(Tabpage tabpage, Error *err)
   try_end(err);
 }
 
+void vim_subscribe(uint64_t channel_id, String event)
+{
+  size_t length = (event.size < EVENT_MAXLEN ? event.size : EVENT_MAXLEN);
+  char e[EVENT_MAXLEN + 1];
+  memcpy(e, event.data, length);
+  e[length] = NUL;
+  channel_subscribe(channel_id, e);
+}
+
+void vim_unsubscribe(uint64_t channel_id, String event)
+{
+  size_t length = (event.size < EVENT_MAXLEN ? event.size : EVENT_MAXLEN);
+  char e[EVENT_MAXLEN + 1];
+  memcpy(e, event.data, length);
+  e[length] = NUL;
+  channel_unsubscribe(channel_id, e);
+}
+
 static void write_msg(String message, bool to_err)
 {
   static int pos = 0;

src/nvim/api/vim.h

@@ -155,5 +155,17 @@ Tabpage vim_get_current_tabpage(void);
 /// @param[out] err Details of an error that may have occurred
 void vim_set_current_tabpage(Tabpage tabpage, Error *err);
 
+/// Subscribes to event broadcasts
+///
+/// @param channel_id The channel id(passed automatically by the dispatcher)
+/// @param event The event type string
+void vim_subscribe(uint64_t channel_id, String event);
+
+/// Unsubscribes to event broadcasts
+///
+/// @param channel_id The channel id(passed automatically by the dispatcher)
+/// @param event The event type string
+void vim_unsubscribe(uint64_t channel_id, String event);
+
 #endif  // NVIM_API_VIM_H
 

src/nvim/eval.c

@@ -13069,7 +13069,7 @@ static void f_send_event(typval_T *argvars, typval_T *rettv)
     return;
   }
 
-  if (argvars[0].v_type != VAR_NUMBER || argvars[0].vval.v_number <= 0) {
+  if (argvars[0].v_type != VAR_NUMBER || argvars[0].vval.v_number < 0) {
     EMSG2(_(e_invarg2), "Channel id must be a positive integer");
     return;
   }

src/nvim/lib/kvec.h

@@ -0,0 +1,91 @@
+/* The MIT License
+
+   Copyright (c) 2008, by Attractive Chaos <attractor@live.co.uk>
+
+   Permission is hereby granted, free of charge, to any person obtaining
+   a copy of this software and associated documentation files (the
+   "Software"), to deal in the Software without restriction, including
+   without limitation the rights to use, copy, modify, merge, publish,
+   distribute, sublicense, and/or sell copies of the Software, and to
+   permit persons to whom the Software is furnished to do so, subject to
+   the following conditions:
+
+   The above copyright notice and this permission notice shall be
+   included in all copies or substantial portions of the Software.
+
+   THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+   EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+   MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+   NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+   BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+   ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+   CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+   SOFTWARE.
+*/
+
+/*
+  An example:
+
+#include "kvec.h"
+int main() {
+	kvec_t(int) array;
+	kv_init(array);
+	kv_push(int, array, 10); // append
+	kv_a(int, array, 20) = 5; // dynamic
+	kv_A(array, 20) = 4; // static
+	kv_destroy(array);
+	return 0;
+}
+*/
+
+/*
+  2008-09-22 (0.1.0):
+
+	* The initial version.
+
+*/
+
+#ifndef AC_KVEC_H
+#define AC_KVEC_H
+
+#include <stdlib.h>
+#include "nvim/memory.h"
+
+#define kv_roundup32(x) (--(x), (x)|=(x)>>1, (x)|=(x)>>2, (x)|=(x)>>4, (x)|=(x)>>8, (x)|=(x)>>16, ++(x))
+
+#define kvec_t(type) struct { size_t n, m; type *a; }
+#define kv_init(v) ((v).n = (v).m = 0, (v).a = 0)
+#define kv_destroy(v) free((v).a)
+#define kv_A(v, i) ((v).a[(i)])
+#define kv_pop(v) ((v).a[--(v).n])
+#define kv_size(v) ((v).n)
+#define kv_max(v) ((v).m)
+
+#define kv_resize(type, v, s)  ((v).m = (s), (v).a = (type*)xrealloc((v).a, sizeof(type) * (v).m))
+
+#define kv_copy(type, v1, v0) do {							\
+		if ((v1).m < (v0).n) kv_resize(type, v1, (v0).n);	\
+		(v1).n = (v0).n;									\
+		memcpy((v1).a, (v0).a, sizeof(type) * (v0).n);		\
+	} while (0)												\
+
+#define kv_push(type, v, x) do {									\
+		if ((v).n == (v).m) {										\
+			(v).m = (v).m? (v).m<<1 : 2;							\
+			(v).a = (type*)xrealloc((v).a, sizeof(type) * (v).m);	\
+		}															\
+		(v).a[(v).n++] = (x);										\
+	} while (0)
+
+#define kv_pushp(type, v) (((v).n == (v).m)?							\
+						   ((v).m = ((v).m? (v).m<<1 : 2),				\
+							(v).a = (type*)xrealloc((v).a, sizeof(type) * (v).m), 0)	\
+						   : 0), ((v).a + ((v).n++))
+
+#define kv_a(type, v, i) (((v).m <= (size_t)(i)? \
+						  ((v).m = (v).n = (i) + 1, kv_roundup32((v).m), \
+						   (v).a = (type*)xrealloc((v).a, sizeof(type) * (v).m), 0) \
+						  : (v).n <= (size_t)(i)? (v).n = (i) + 1 \
+						  : 0), (v).a[(i)])
+
+#endif

src/nvim/os/channel.c

@@ -15,9 +15,11 @@
 #include "nvim/vim.h"
 #include "nvim/memory.h"
 #include "nvim/map.h"
+#include "nvim/lib/kvec.h"
 
 typedef struct {
   uint64_t id;
+  Map(cstr_t) *subscribed_events;
   bool is_job;
   msgpack_unpacker *unpacker;
   msgpack_sbuffer *sbuffer;
@@ -33,17 +35,24 @@ typedef struct {
 
 static uint64_t next_id = 1;
 static Map(uint64_t) *channels = NULL;
+static Map(cstr_t) *event_strings = NULL;
 static msgpack_sbuffer msgpack_event_buffer;
 
-static void on_job_stdout(RStream *rstream, void *data, bool eof);
-static void on_job_stderr(RStream *rstream, void *data, bool eof);
+static void job_out(RStream *rstream, void *data, bool eof);
+static void job_err(RStream *rstream, void *data, bool eof);
 static void parse_msgpack(RStream *rstream, void *data, bool eof);
+static void send_event(Channel *channel, char *type, typval_T *data);
+static void broadcast_event(char *type, typval_T *data);
+static void unsubscribe(Channel *channel, char *event);
 static void close_channel(Channel *channel);
 static void close_cb(uv_handle_t *handle);
+static WBuffer *serialize_event(char *type, typval_T *data);
+static Channel *register_channel(void);
 
 void channel_init()
 {
   channels = map_new(uint64_t)();
+  event_strings = map_new(cstr_t)();
   msgpack_sbuffer_init(&msgpack_event_buffer);
 }
 
@@ -62,71 +71,78 @@ void channel_teardown()
 
 void channel_from_job(char **argv)
 {
-  Channel *channel = xmalloc(sizeof(Channel));
-  rstream_cb rcb = on_job_stdout;
-  channel->unpacker = msgpack_unpacker_new(MSGPACK_UNPACKER_INIT_BUFFER_SIZE);
-  channel->sbuffer = msgpack_sbuffer_new();
-
-  channel->id = next_id++;
+  Channel *channel = register_channel();
   channel->is_job = true;
-  channel->data.job_id = job_start(argv, channel, rcb, on_job_stderr, NULL);
-  map_put(uint64_t)(channels, channel->id, channel);
+  channel->data.job_id = job_start(argv, channel, job_out, job_err, NULL);
 }
 
 void channel_from_stream(uv_stream_t *stream)
 {
-  Channel *channel = xmalloc(sizeof(Channel));
-  rstream_cb rcb = parse_msgpack;
-  channel->unpacker = msgpack_unpacker_new(MSGPACK_UNPACKER_INIT_BUFFER_SIZE);
-  channel->sbuffer = msgpack_sbuffer_new();
-
+  Channel *channel = register_channel();
   stream->data = NULL;
-  channel->id = next_id++;
   channel->is_job = false;
   // read stream
-  channel->data.streams.read = rstream_new(rcb, 1024, channel, true);
+  channel->data.streams.read = rstream_new(parse_msgpack, 1024, channel, true);
   rstream_set_stream(channel->data.streams.read, stream);
   rstream_start(channel->data.streams.read);
   // write stream
   channel->data.streams.write = wstream_new(1024 * 1024);
   wstream_set_stream(channel->data.streams.write, stream);
   channel->data.streams.uv = stream;
-  map_put(uint64_t)(channels, channel->id, channel);
 }
 
 bool channel_send_event(uint64_t id, char *type, typval_T *data)
 {
-  Channel *channel = map_get(uint64_t)(channels, id);
+  Channel *channel = NULL;
 
-  if (!channel) {
-    return false;
+  if (id > 0) {
+    if (!(channel = map_get(uint64_t)(channels, id))) {
+      return false;
+    }
+    send_event(channel, type, data);
+  } else {
+    broadcast_event(type, data);
   }
 
-  String event_type = {.size = strnlen(type, 1024), .data = type};
-  Object event_data = vim_to_object(data);
-  msgpack_packer packer;
-  msgpack_packer_init(&packer, &msgpack_event_buffer, msgpack_sbuffer_write);
-  msgpack_rpc_notification(event_type, event_data, &packer);
-  char *bytes = xmemdup(msgpack_event_buffer.data, msgpack_event_buffer.size);
-
-  wstream_write(channel->data.streams.write,
-                bytes,
-                msgpack_event_buffer.size,
-                true);
-
-  msgpack_rpc_free_object(event_data);
-  msgpack_sbuffer_clear(&msgpack_event_buffer);
-
   return true;
 }
 
-static void on_job_stdout(RStream *rstream, void *data, bool eof)
+void channel_subscribe(uint64_t id, char *event)
+{
+  Channel *channel;
+
+  if (!(channel = map_get(uint64_t)(channels, id))) {
+    return;
+  }
+
+  char *event_string = map_get(cstr_t)(event_strings, event);
+
+  if (!event_string) {
+    event_string = xstrdup(event);
+    map_put(cstr_t)(event_strings, event_string, event_string);
+  }
+
+  map_put(cstr_t)(channel->subscribed_events, event_string, event_string);
+}
+
+void channel_unsubscribe(uint64_t id, char *event)
+{
+  Channel *channel;
+
+  if (!(channel = map_get(uint64_t)(channels, id))) {
+    return;
+  }
+
+  unsubscribe(channel, event);
+}
+
+static void job_out(RStream *rstream, void *data, bool eof)
 {
   Job *job = data;
   parse_msgpack(rstream, job_data(job), eof);
 }
 
-static void on_job_stderr(RStream *rstream, void *data, bool eof)
+static void job_err(RStream *rstream, void *data, bool eof)
 {
   // TODO(tarruda): plugin error messages should be sent to the error buffer
 }
@@ -158,15 +174,62 @@ static void parse_msgpack(RStream *rstream, void *data, bool eof)
     // Perform the call
     msgpack_rpc_call(channel->id, &unpacked.data, &response);
     wstream_write(channel->data.streams.write,
-                  xmemdup(channel->sbuffer->data, channel->sbuffer->size),
-                  channel->sbuffer->size,
-                  true);
+                  wstream_new_buffer(channel->sbuffer->data,
+                                     channel->sbuffer->size,
+                                     true));
 
     // Clear the buffer for future calls
     msgpack_sbuffer_clear(channel->sbuffer);
   }
 }
 
+static void send_event(Channel *channel, char *type, typval_T *data)
+{
+  wstream_write(channel->data.streams.write, serialize_event(type, data));
+}
+
+static void broadcast_event(char *type, typval_T *data)
+{
+  kvec_t(Channel *) subscribed;
+  kv_init(subscribed);
+  Channel *channel;
+
+  map_foreach_value(channels, channel, {
+    if (map_has(cstr_t)(channel->subscribed_events, type)) {
+      kv_push(Channel *, subscribed, channel);
+    }
+  });
+
+  if (!kv_size(subscribed)) {
+    goto end;
+  }
+
+  WBuffer *buffer = serialize_event(type, data);
+
+  for (size_t i = 0; i < kv_size(subscribed); i++) {
+    wstream_write(kv_A(subscribed, i)->data.streams.write, buffer);
+  }
+
+end:
+  kv_destroy(subscribed);
+}
+
+static void unsubscribe(Channel *channel, char *event)
+{
+  char *event_string = map_get(cstr_t)(event_strings, event);
+  map_del(cstr_t)(channel->subscribed_events, event_string);
+
+  map_foreach_value(channels, channel, {
+    if (map_has(cstr_t)(channel->subscribed_events, event_string)) {
+      return;
+    }
+  });
+
+  // Since the string is no longer used by other channels, release it's memory
+  map_del(cstr_t)(event_strings, event_string);
+  free(event_string);
+}
+
 static void close_channel(Channel *channel)
 {
   map_del(uint64_t)(channels, channel->id);
@@ -181,6 +244,13 @@ static void close_channel(Channel *channel)
     uv_close((uv_handle_t *)channel->data.streams.uv, close_cb);
   }
 
+  // Unsubscribe from all events
+  char *event_string;
+  map_foreach_value(channel->subscribed_events, event_string, {
+    unsubscribe(channel, event_string);
+  });
+
+  map_free(cstr_t)(channel->subscribed_events);
   free(channel);
 }
 
@@ -190,3 +260,29 @@ static void close_cb(uv_handle_t *handle)
   free(handle);
 }
 
+static WBuffer *serialize_event(char *type, typval_T *data)
+{
+  String event_type = {.size = strnlen(type, EVENT_MAXLEN), .data = type};
+  Object event_data = vim_to_object(data);
+  msgpack_packer packer;
+  msgpack_packer_init(&packer, &msgpack_event_buffer, msgpack_sbuffer_write);
+  msgpack_rpc_notification(event_type, event_data, &packer);
+  WBuffer *rv = wstream_new_buffer(msgpack_event_buffer.data,
+                                   msgpack_event_buffer.size,
+                                   true);
+  msgpack_rpc_free_object(event_data);
+  msgpack_sbuffer_clear(&msgpack_event_buffer);
+
+  return rv;
+}
+
+static Channel *register_channel()
+{
+  Channel *rv = xmalloc(sizeof(Channel));
+  rv->unpacker = msgpack_unpacker_new(MSGPACK_UNPACKER_INIT_BUFFER_SIZE);
+  rv->sbuffer = msgpack_sbuffer_new();
+  rv->id = next_id++;
+  rv->subscribed_events = map_new(cstr_t)();
+  map_put(uint64_t)(channels, rv->id, rv);
+  return rv;
+}

src/nvim/os/channel.h

@@ -5,6 +5,8 @@
 
 #include "nvim/vim.h"
 
+#define EVENT_MAXLEN 512
+
 /// Initializes the module
 void channel_init(void);
 
@@ -25,11 +27,24 @@ void channel_from_job(char **argv);
 
 /// Sends event/data to channel
 ///
-/// @param id The channel id
+/// @param id The channel id. If 0, the event will be sent to all
+///        channels that have subscribed to the event type
 /// @param type The event type, an arbitrary string
 /// @param obj The event data
 /// @return True if the data was sent successfully, false otherwise.
 bool channel_send_event(uint64_t id, char *type, typval_T *data);
 
+/// Subscribes to event broadcasts
+///
+/// @param id The channel id
+/// @param event The event type string
+void channel_subscribe(uint64_t id, char *event);
+
+/// Unsubscribes to event broadcasts
+///
+/// @param id The channel id
+/// @param event The event type string
+void channel_unsubscribe(uint64_t id, char *event);
+
 #endif  // NVIM_OS_CHANNEL_H
 

src/nvim/os/job.c

@@ -234,7 +234,7 @@ bool job_write(int id, char *data, uint32_t len)
     return false;
   }
 
-  if (!wstream_write(job->in, data, len, true)) {
+  if (!wstream_write(job->in, wstream_new_buffer(data, len, false))) {
     job_stop(job->id);
     return false;
   }

src/nvim/os/rstream.c

@@ -33,7 +33,7 @@ static void close_cb(uv_handle_t *handle);
 static void emit_read_event(RStream *rstream, bool eof);
 
 RStream * rstream_new(rstream_cb cb,
-                      uint32_t buffer_size,
+                      size_t buffer_size,
                       void *data,
                       bool async)
 {
@@ -133,7 +133,7 @@ void rstream_stop(RStream *rstream)
   }
 }
 
-size_t rstream_read(RStream *rstream, char *buf, uint32_t count)
+size_t rstream_read(RStream *rstream, char *buf, size_t count)
 {
   size_t read_count = rstream->wpos - rstream->rpos;
 

src/nvim/os/rstream.h

@@ -21,7 +21,7 @@
 ///        this to false
 /// @return The newly-allocated `RStream` instance
 RStream * rstream_new(rstream_cb cb,
-                      uint32_t buffer_size,
+                      size_t buffer_size,
                       void *data,
                       bool async);
 
@@ -65,7 +65,7 @@ void rstream_stop(RStream *rstream);
 /// @param buffer The buffer which will receive the data
 /// @param count Number of bytes that `buffer` can accept
 /// @return The number of bytes copied into `buffer`
-size_t rstream_read(RStream *rstream, char *buffer, uint32_t count);
+size_t rstream_read(RStream *rstream, char *buffer, size_t count);
 
 /// Returns the number of bytes available for reading from `rstream`
 ///

src/nvim/os/wstream.c

@@ -12,27 +12,27 @@
 struct wstream {
   uv_stream_t *stream;
   // Memory currently used by pending buffers
-  uint32_t curmem;
+  size_t curmem;
   // Maximum memory used by this instance
-  uint32_t maxmem;
+  size_t maxmem;
   // Number of pending requests
-  uint32_t pending_reqs;
+  size_t pending_reqs;
   bool freed;
 };
 
+struct wbuffer {
+  size_t refcount, size;
+  char *data;
+};
+
 typedef struct {
   WStream *wstream;
-  // Buffer containing data to be written
-  char *buffer;
-  // Size of the buffer
-  uint32_t length;
-  // If it's our responsibility to free the buffer
-  bool free;
+  WBuffer *buffer;
 } WriteData;
 
 static void write_cb(uv_write_t *req, int status);
 
-WStream * wstream_new(uint32_t maxmem)
+WStream * wstream_new(size_t maxmem)
 {
   WStream *rv = xmalloc(sizeof(WStream));
   rv->maxmem = maxmem;
@@ -59,51 +59,60 @@ void wstream_set_stream(WStream *wstream, uv_stream_t *stream)
   wstream->stream = stream;
 }
 
-bool wstream_write(WStream *wstream, char *buffer, uint32_t length, bool free)
+bool wstream_write(WStream *wstream, WBuffer *buffer)
 {
   WriteData *data;
   uv_buf_t uvbuf;
   uv_write_t *req;
 
-  if (wstream->freed) {
-    // Don't accept write requests after the WStream instance was freed
+  // This should not be called after a wstream was freed
+  assert(!wstream->freed);
+
+  if (wstream->curmem + buffer->size > wstream->maxmem) {
     return false;
   }
 
-  if (wstream->curmem + length > wstream->maxmem) {
-    return false;
-  }
-
-  if (free) {
-    // We should only account for buffers that are ours to free
-    wstream->curmem += length;
-  }
-
+  buffer->refcount++;
+  wstream->curmem += buffer->size;
   data = xmalloc(sizeof(WriteData));
   data->wstream = wstream;
   data->buffer = buffer;
-  data->length = length;
-  data->free = free;
   req = xmalloc(sizeof(uv_write_t));
   req->data = data;
-  uvbuf.base = buffer;
-  uvbuf.len = length;
+  uvbuf.base = buffer->data;
+  uvbuf.len = buffer->size;
   wstream->pending_reqs++;
   uv_write(req, wstream->stream, &uvbuf, 1, write_cb);
 
   return true;
 }
 
+WBuffer *wstream_new_buffer(char *data, size_t size, bool copy)
+{
+  WBuffer *rv = xmalloc(sizeof(WBuffer));
+  rv->size = size;
+  rv->refcount = 0;
+
+  if (copy) {
+    rv->data = xmemdup(data, size);
+  } else {
+    rv->data = data;
+  }
+
+  return rv;
+}
+
 static void write_cb(uv_write_t *req, int status)
 {
   WriteData *data = req->data;
 
   free(req);
+  data->wstream->curmem -= data->buffer->size;
 
-  if (data->free) {
+  if (!--data->buffer->refcount) {
     // Free the data written to the stream
+    free(data->buffer->data);
     free(data->buffer);
-    data->wstream->curmem -= data->length;
   }
 
   data->wstream->pending_reqs--;

src/nvim/os/wstream.h

@@ -12,7 +12,7 @@
 ///
 /// @param maxmem Maximum amount memory used by this `WStream` instance.
 /// @return The newly-allocated `WStream` instance
-WStream * wstream_new(uint32_t maxmem);
+WStream * wstream_new(size_t maxmem);
 
 /// Frees all memory allocated for a WStream instance
 ///
@@ -31,10 +31,19 @@ void wstream_set_stream(WStream *wstream, uv_stream_t *stream);
 ///
 /// @param wstream The `WStream` instance
 /// @param buffer The buffer which contains data to be written
-/// @param length Number of bytes that should be written from `buffer`
-/// @param free If true, `buffer` will be freed after the write is complete
-/// @return true if the data was successfully queued, false otherwise.
-bool wstream_write(WStream *wstream, char *buffer, uint32_t length, bool free);
+/// @return false if the write failed
+bool wstream_write(WStream *wstream, WBuffer *buffer);
+
+/// Creates a WBuffer object for holding output data. Instances of this
+/// object can be reused across WStream instances, and the memory is freed
+/// automatically when no longer needed(it tracks the number of references
+/// internally)
+///
+/// @param data Data stored by the WBuffer
+/// @param size The size of the data array
+/// @param copy If true, the data will be copied into the WBuffer
+/// @return The allocated WBuffer instance
+WBuffer *wstream_new_buffer(char *data, size_t size, bool copy);
 
 #endif  // NVIM_OS_WSTREAM_H
 

src/nvim/os/wstream_defs.h

@@ -1,6 +1,7 @@
 #ifndef NVIM_OS_WSTREAM_DEFS_H
 #define NVIM_OS_WSTREAM_DEFS_H
 
+typedef struct wbuffer WBuffer;
 typedef struct wstream WStream;
 
 #endif  // NVIM_OS_WSTREAM_DEFS_H