Performance regression compiling glium

[XMPPwocky]

http://github.com/tomaka/glium

Used to take <10 minutes to compile in release builds, but with nightly (5/15/15) I'm sitting at around the 2-hour mark and it's still not done building.

[shepmaster]

Perhaps related to #25393?

[Ryman]

I think this was fixed by the related issue? Otherwise we would probably still be hearing about it?

[MagaTailor]

Seeing as this issue was never closed, glium currently takes a few hours (never timed it, but probably around 4-5) on my ARM machine whereas in the past it used to take, IIRC, 10-15 or maybe even fewer minutes. Seems like a huge regression.

[arielb1]

Are you using -Z orbit? Could you give us the output of -Z time-passes?

[MagaTailor]

I'll have to leave it running overnight. And yes, I default to -Z orbit.

glium-0.13.5_time_flies.txt from rustc 1.11.0-dev (8a50e295f 2016-07-02)

[nikomatsakis]

Is this still an issue?

[eddyb]

@petevine Do you have the timings without -Z orbit? The issue is older than MIR trans.

[MagaTailor]

It's the same issue as the linked one, I already OK'ed @jschievink's findings. (compilation time back to normal)

[eddyb]

Alright then, the original issue appears to have been fixed long ago.