04:14 <MajorLag> I expect to be able to push the Stream portion of interface reform when I wake up tomorrow morning (running what should be final integration tests now). That just leaves Random, which tiehuis said he already did.

04:20 <MajorLag> There was a little bit of annoyance from the Abstract interfaces requiring a *@OpaqueType, because some implementations didn't need to be a pointer and a couple were 0-bit types which I had to add a dummy field to.

04:21 <MajorLag> Of course, they needed an interface struct member before, so I guess nothing was really lost there.

16:17 <MajorLag> Is godbolt broken for anyone else? Seems to insist on inserting \r\n newlines.

16:20 <MajorLag> Oh, I see that issue was reported 2 days ago

16:40 <forgot-password> Is a inline for loop always run at compile time?

16:40 <forgot-password> Or can I call runtime functions inside of it?

16:51 <MajorLag> you can call runtime functions inside it, but whatever it is iterating over must be comptime known.

17:09 <forgot-password> Alright, so then my loop just gets unrolled into multiple calls, which are populated with the comptime values from the iteration?

17:10 <MajorLag> yes

17:54 <forgot-p1ssword> Is there any plan on supporting information being carried with an error?

17:54 <forgot-p1ssword> Currently I'm using std.debug.warn to print some text and then return an error

17:59 <MajorLag> andrewrk, I have a question when you have time. Given this simplified gist: https://gist.github.com/tgschultz/c9033629fa1625fd166b8a38c5f9c4f6, at first glance it looks like the abstract stream shouldn't work, since it passes `*File` where `File` is expected, however it seems to work fine. I suspect this is because `File` is really `*const File` under the hood. Can you confirm that I shouldn't rely on this behavior

17:59 <MajorLag> p being passed as an actual copy at some point in the future?

20:11 <andrewrk> MajorLag, I think your code isn't being analyzed because it's not being called. lazy top level declaration evaluation

21:12 <MajorLag> andrewrk, it is being analyzed, I have it called by parts of std and passing a full integration test. I think I'm just getting lucky because passing a struct as a parameter currently means it is always passed as a pointer (IIRC), but that will change in the future and the compiler could choose to pass it as a copy and break it.

21:14 <andrewrk> that's correct