# haxe-ui
- yeah, im seeing the exact same thingb
bright-gpu-74537
04/10/2023, 2:28 PMyeah, im seeing the exact same thing - nice, i thought i was doing something silly 😄a
ambitious-knife-25690
04/10/2023, 2:28 PMnice, i thought i was doing something silly 😄 - it is more "reliable" with a timer, in that all the requests go through but it seems like timer + threaded http call does not work together *at all*a
ambitious-knife-25690
04/10/2023, 2:29 PMit is more "reliable" with a timer, in that all the requests go through but it seems like timer + threaded http call does not work together at all - yeah, i mean, this _could_ all be because of the threading, but it seems bizarrely consistentb
bright-gpu-74537
04/10/2023, 2:30 PMyeah, i mean, this could all be because of the threading, but it seems bizarrely consistent - it isa
ambitious-knife-25690
04/10/2023, 2:30 PMit is - i had assumed it was something to do with the data i was getting back from the sitea
ambitious-knife-25690
04/10/2023, 2:30 PMi had assumed it was something to do with the data i was getting back from the site - but the data seemed to all be "okay" and nothing seemed to be emptya
ambitious-knife-25690
04/10/2023, 2:30 PMbut the data seemed to all be "okay" and nothing seemed to be empty - at least on first glancea
ambitious-knife-25690
04/10/2023, 2:31 PMat least on first glance - mine consistently freezes on the 4tha
ambitious-knife-25690
04/10/2023, 2:31 PMmine consistently freezes on the 4th - Does this only occur in a HaxeUI app, or can you reproduce it in a standalone test case?p
powerful-morning-89
04/10/2023, 2:40 PMDoes this only occur in a HaxeUI app, or can you reproduce it in a standalone test case? - not sure yet, currently its a haxeui app, ill try to repro it in a non haxeui app when i get a little more understanding of whats going onb
bright-gpu-74537
04/10/2023, 2:41 PMnot sure yet, currently its a haxeui app, ill try to repro it in a non haxeui app when i get a little more understanding of whats going on - this is the line it freezes on: `api.writeBytes(buf, 0, len);`b
bright-gpu-74537
04/10/2023, 2:42 PMthis is the line it freezes on:api.writeBytes(buf, 0, len); - Hmm. I seem to remember running into weird freezing stuff with sockets when I was trying to make an async socket...p
powerful-morning-89
04/10/2023, 2:43 PMHmm. I seem to remember running into weird freezing stuff with sockets when I was trying to make an async socket... - sounds pretty similarb
bright-gpu-74537
04/10/2023, 2:46 PMsounds pretty similar - But iirc it was a bug in `haxe.io.Input`'s readAll implementation, not in writeBytes.p
powerful-morning-89
04/10/2023, 2:50 PMBut iirc it was a bug in `haxe.io.Input`'s readAll implementation, not in writeBytes. - right, defo could be that here also i think... not sure... it does seem to get bytes, but then hangs somewhere after that... really hard to debug this toob
bright-gpu-74537
04/10/2023, 2:51 PMright, defo could be that here also i think... not sure... it does seem to get bytes, but then hangs somewhere after that... really hard to debug this too - ```haxe trace("11111111199999999999999999999", pos, len, pos + len); for (i in pos...pos + len) b.push(b2[i]); trace("22222222222222299999999999999999999"); ```b
bright-gpu-74537
04/10/2023, 3:07 PMCopy codehaxe trace("11111111199999999999999999999", pos, len, pos + len); for (i in pos...pos + len) b.push(b2[i]); trace("22222222222222299999999999999999999"); - the second trace never appears (from BytesBuffer.hx)b
bright-gpu-74537
04/10/2023, 3:07 PMthe second trace never appears (from BytesBuffer.hx) - I would suggest using a debugger and digging deeper. (A real debugger, not that hxcpp thing that'll give up the moment you reach native code.)p
powerful-morning-89
04/10/2023, 3:12 PMI would suggest using a debugger and digging deeper. (A real debugger, not that hxcpp thing that'll give up the moment you reach native code.) - yeah, might what it through visual studio, see what (if anything) it comes up withb
bright-gpu-74537
04/10/2023, 3:13 PMyeah, might what it through visual studio, see what (if anything) it comes up with - so, i actually think its this code in core/http: https://github.com/core-haxe/http/blob/main/src/http/providers/DefaultHttpProvider.hx#L28-L44b
bright-gpu-74537
04/10/2023, 3:51 PMso, i actually think its this code in core/http: https://github.com/core-haxe/http/blob/main/src/http/providers/DefaultHttpProvider.hx#L28-L44 - i can make it work _alot_ more consistently by NOT calling resolve from the thread and instead having a timer pull "resolves" off of an array and call them from the main threadb
bright-gpu-74537
04/10/2023, 3:51 PMi can make it work alot more consistently by NOT calling resolve from the thread and instead having a timer pull "resolves" off of an array and call them from the main thread - b
bright-gpu-74537
04/10/2023, 3:51 PMhttps://cdn.discordapp.com/attachments/565569107701923852/1095013297952149514/image.pngâ–¾
- (first time ive seen that)b
bright-gpu-74537
04/10/2023, 3:52 PM(first time ive seen that) - You indeed shouldn't call callbacks on the worker thread.p
powerful-morning-89
04/10/2023, 3:53 PMYou indeed shouldn't call callbacks on the worker thread. - yeah, feels obvious now... i threw that together in 10 mins... should have thought about it a little more and wouldnt have spend the day debugging something pretty simpleb
bright-gpu-74537
04/10/2023, 3:54 PMyeah, feels obvious now... i threw that together in 10 mins... should have thought about it a little more and wouldnt have spend the day debugging something pretty simple - b
bright-gpu-74537
04/10/2023, 3:54 PMhttps://cdn.discordapp.com/attachments/565569107701923852/1095014035965091930/image.pngâ–¾
- very consistently "right" nowb
bright-gpu-74537
04/10/2023, 3:55 PMvery consistently "right" now - > basic threaded request, which allows for async programming, could / should be greatly > improved by using a thread pool, but for a preliminary impl its better than the > standard (sync) behaviour of haxe std http Note that there's a thread pool implementation in the std.p
powerful-morning-89
04/10/2023, 3:56 PM> basic threaded request, which allows for async programming, could / should be greatly > improved by using a thread pool, but for a preliminary impl its better than the > standard (sync) behaviour of haxe std http Note that there's a thread pool implementation in the std. - oh yeah, forgot about that... cool - will check it out... at the moment i just wanna fix the bad impl that is there - cheers for the headsupb
bright-gpu-74537
04/10/2023, 3:56 PMoh yeah, forgot about that... cool - will check it out... at the moment i just wanna fix the bad impl that is there - cheers for the headsup