New Compiler

Hi Jared,

We've been debugging a similar issue here all day. Everything was working fine last night, but since this morning there's been an odd phenomenon where each project seems to be getting into an inconsistent build state as soon as *any* code in the project causes a build timeout (and build timeouts seem to occur as soon as a given file gets to around ~256 or so lines).

After the first build timeout within a project, nothing we try to compile in that same project ever works again, regardless of what code we're trying to run. So, for example, if we:

(1) Create a new project, and build the "Basic Template Algorithm" code as is without changing anything, it works.

(2) We then delete the template code, and in its place we copy/paste some code that compiled just fine last night from a local backup. The file is about 500 lines, but it could be any length up to around ~256 or so. That will then cause a build timeout.

(3) We then hit Ctrl+Z until the project window contains the original ""Basic Template Algorithm"" code, and now that template code that just worked a minute ago will fail to build indefinitely, no matter how many times we try to rebuild it. 

So in summary, we can create a situation where any given code fragment that compiles successfully in attempt (1) will then fail to compile in attempts (3) - (N), if the code we run in step (2) happens to cause a build timeout.

Mentioning this here because it sounds like it's likely to be an issue with this stateful compilation business.

Any ideas on how we can get around this?

Thanks. :)

Hey Jason! If possible when you find an edge case like that please send a ticket via the support tab in the left of the IDE with the project attached so we can repeat/debug it.

Tomac; we found a bug where it's not reporting illegal function usage; if you rename your method to not look like "open(" it'll work.

Thanks Jared, will do! Just to clarify though, I didn't mean it to sound like an edge case. What I should have said is that we can't run anything on the platform anymore (since yesterday morning), and every project that was previously working is now stuck in a build timeout loop, which sounds similar to the issues that other folks in the thread are facing, and it seems like it might be related to the stateful compilation. The post from before was just my (probably a bit too detailed) attempt to describe the failure mode we were seeing, so that it might help with pinpointing the source of the problem. I'll go submit a support ticket too, but I'd be interested to know if anyone else is running into similar problems. Anyways, thanks again for the help, as always. :)

Hi Jared, Thanks for the feedback. But I am having the same experience as Jason Wilkes. Effectively, anything I want to build gets the "Build Timeout, please check your internet connection." msg. Do you see similar experience on your end? 

Hi Tomac every issue is unique; please send a repeatable case to support so we can debug it via the left support tab or support@quantconnect.com.

BTW Tomac I already replied to you above:
'Tomac; we found a bug where it's not reporting illegal function usage; if you rename your method to not look like "open(" it'll work."

Wow - the functionality on this site has really grown since the last time I was here.  Love the Bootcamp feature - looks very similar to the interface used on datacamp.com which I use almost daily.

I'm impressed! I'll have backtest up and running very soon. ty

Quick question please.

Will this platform allow me to mix and match within the same strategy different timeframes ?

i.e Place a market order when the SMA/Hourly crossess over the EMA/15min  ?

ty in advance.

Yes, the mix and match of different timeframes within the same algorithm can be done via consolidators. These consolidators allow users to consolidate smaller data points into larger bars, and then these larger bars can be fed to indicators. In the algorithm below, trades are made when a 15 minute EMA crosses over an hourly SMA:

Great to know. Thank you for the quick follow up!