Move more logic into the reader. Avoid copying partial messages in
cases where we know that the buffer is large enough.
This is mostly groundwork for trying to add support for continuation
frames.
Adding HTTP & websocket awareness to the TCP server.
HTTP server handles `GET /json/version` and websocket upgrade requests.
Conceptually, websocket handling is the same code as before, but receiving
data will parse the websocket frames and writing data will wrap it in
a websocket frame.
The previous `Ctx` was split into a `Server` and a `Client`. This was
largely done to make it easy to write unit tests, since the `Client` is
a generic, all its dependencies (i.e. the server) can be mocked out. This
also makes it a bit nicer to know if there is or isn't a client (via the
server's client optional).
Added a MemoryPool for the Send object (I thought that was a nice touch!)
Removed MacOS hack on accept/conn completion usage.
Known issues:
- When framing an outgoing message, the entire message has to be duped. This
is no worse than how it was before, but it should be possible to eliminate
this in the future. Probably not part of this PR.
- Websocket parsing will reject continuation frames. I don't know of a single
client that will send a fragmented message (websocket has its own
message fragmentation), but we should probably still support this just in
case.
- I don't think the receive, timeout and close completions can safely be
re-used like we're doing. I believe they need to be associated with a specific
client socket.
- A new connection creates a new browser session. I think this is right (??),
but for the very first, we're throwing out a perfectly usable session. I'm
thinking this might be a change to how Browser/Sessions work.
- zig build test won't compile. This branch reproduces the issue with none
of these changes:
https://github.com/karlseguin/browser/tree/broken_test_build
(or, as a diff to main):
https://github.com/lightpanda-io/browser/compare/main...karlseguin:broken_test_build
CDP is now an struct which contains its own state a browser and a session.
When a client connection is made and successfully upgrades, the client creates
the CDP instance. There is now a cleaner separation betwen Server, Client and
CDP.
Removed a number of allocations, especially when writing results/events from
CDP to the client. Improved input message parsing. Tried to remove some usage
of undefined.
Move more logic into the reader. Avoid copying partial messages in
cases where we know that the buffer is large enough.
This is mostly groundwork for trying to add support for continuation
frames.
Adding HTTP & websocket awareness to the TCP server.
HTTP server handles `GET /json/version` and websocket upgrade requests.
Conceptually, websocket handling is the same code as before, but receiving
data will parse the websocket frames and writing data will wrap it in
a websocket frame.
The previous `Ctx` was split into a `Server` and a `Client`. This was
largely done to make it easy to write unit tests, since the `Client` is
a generic, all its dependencies (i.e. the server) can be mocked out. This
also makes it a bit nicer to know if there is or isn't a client (via the
server's client optional).
Added a MemoryPool for the Send object (I thought that was a nice touch!)
Removed MacOS hack on accept/conn completion usage.
Known issues:
- When framing an outgoing message, the entire message has to be duped. This
is no worse than how it was before, but it should be possible to eliminate
this in the future. Probably not part of this PR.
- Websocket parsing will reject continuation frames. I don't know of a single
client that will send a fragmented message (websocket has its own
message fragmentation), but we should probably still support this just in
case.
- I don't think the receive, timeout and close completions can safely be
re-used like we're doing. I believe they need to be associated with a specific
client socket.
- A new connection creates a new browser session. I think this is right (??),
but for the very first, we're throwing out a perfectly usable session. I'm
thinking this might be a change to how Browser/Sessions work.
- zig build test won't compile. This branch reproduces the issue with none
of these changes:
https://github.com/karlseguin/browser/tree/broken_test_build
(or, as a diff to main):
https://github.com/lightpanda-io/browser/compare/main...karlseguin:broken_test_build
1 - Use getOrPut to avoid making 2 map lookups where possible.
2 - Use an arena allocator for Values, which makes memory management simpler.
3 - Because of #2, we no longer need to allocate key or values which don't need
to be unescaped. The downside is that the input string has to outlive the
query.Values (but I think this is currently always the case)
4 - Optimize unescape logic & allocations
5 - Improve test coverage
Common cases, text/html, text/xml and text/plain parse about 2x faster. Other
cases are about 30% faster.
Support quoted attributes, i.e. charset="utf-8" & valid escape sequences. This
potentially requires allocation, thus Mime.parse now takes an allocator.
Stricter validation around type/subtype based on RFC.
More tests.
Replace Mime.eql with isHTML(). Equality is complicated and was previously
incorrect (it was case sensitive, it should not be). Since we currently only
use isHTML-like behavior, built a (faster) method specifically for that.
Preserves all existing behavior (i.e. make test and zig build test are not
changed in any way).
The new 'unittest' only runs unit tests and is fast to build. It takes ~1.7 to
build unittest, vs ~11.09 to build test. This is really the main goal, and
hopefully any unit test which are (a) fast and (b) don't impact build times
will be run here.
The test runner is based on:
https://gist.github.com/karlseguin/c6bea5b35e4e8d26af6f81c22cb5d76b
It allow filtering, i.e. `make unittest F="parse query dup"`.
'unittest' does memory leak detection when tests use std.testing.allocator.
Fixed a memory leak in url/query which was detected/reported with by the new
'unittest'.
In order to avoid having 3 src/test_xyx.zig files, I merged the existing
test_runner.zig and run_tests.zig into a single main_tests.zig. (this change
is superfluous, but I thought it was cleaner this way. Happy to revert this).
1- Remove `parser.trim`, it was only being used in 1 place. All other places
are using `std.mem.trim(u8, X, &std.ascii.whitespace)`, so i updated MIME to
use this as well
2- Use slightly more meaningful field name, i => pos, s = data
3- Leverage std.mem.indexOfScalarPos which can be more efficient for longer
inputs (since it leverages SIMD)
Leverage comptime fields to give generated Tuple a default value, allowing
TupleT and Tuple to be merged.
Only call generate.Tuple on the final output. This eliminates redundant
deduplication, and results in a simpler API (nested types just need to expose
a natural Zig tuple).
generate.Union leverages the new Tuple and removes unused features.
