READY | alleviating browser throttling on WebRTC transport #145
Conversation
|
Looks like there's a race condition when ending the stream. |
ya7ya
changed the title
|
@pgte Thanks for the The tests are passing locally just fine. let me know there are any edits required 😄 |
|
BTW. the tests are passing successfully on my machine, I reckon it has to do with the
|
| const block = new BlockStream(this.blockStreamSize, {nopad: true}) | ||
|
|
||
| pump(channel, block, channel, (err) => { | ||
| if (err) return callback(err) |
There was a problem hiding this comment.
I see a possible double-callback here when an error occurs. Maybe use the once package to make sure you only call the callback once?
There was a problem hiding this comment.
@pgte , once is already used few lines before https://github.com/libp2p/js-libp2p-webrtc-star/pull/145/files/1b0d1ed130d6d447c3d0fc54b0abee13fdde11e7#diff-1fdf421c05c1140f6d71444ea2b27638R62
callback = callback ? once(callback) : noop
// ....
// ....
const block = new BlockStream(this.blockStreamSize, {nopad: true})
pump(channel, block, channel, (err) => {
if (err) return callback(err)
})Does that cover it or am i missing something here ?
There was a problem hiding this comment.
Ha, it is, I missed it, sorry. Looks good to me!
| const block = new BlockStream(this.blockStreamSize, {nopad: true}) | ||
|
|
||
| pump(channel, block, channel, (err) => { | ||
| if (err) return callback(err) |
|
I edited the commit msgs that weren't standard. so I had to |
This PR is related to ipfs/js-ipfs#1019 and ipfs/js-ipfs-bitswap#152 , it basically implements
MAX_MESSAGE_SIZEon the webRTC transport to avoid being throttled by the browser.however as you can see, it fails one of the tests. I would appreciate some help in debugging this since It's been frustrating.
I've tried to use
pull-blockbut it didn't work (check the commented out code indial)note this also still requires some clean-up as you might have noticed 😄
cc. @diasdavid