FYI: Another Redmine outage at 1630Z (about 13 minutes from now) as we try to upgrade Redmine itself

*0030Z

t i m e

You mean @62 .beats!

ha!

ooh.. any new features worth noting?

Who can tell, given the amount of Ruby tedium between here and the finish line

hrm.. how upset would people be if cw added '-fdiagnostics-color=always' to gcc runs iff cw's stdout is a terminal? for when you're building interactively (but keep esc codes out of nightly/mail_msg)

(i finally bothered to look, but i'm guessing others already knew) since it saves each compiler's output to a temp file and then writes it out, so it gets disabled

jbk: will it respect NO_COLOR

also, why do you have to force it if output is a terminal

Also, FYI: Redmine is probably good to go now. Let me know if you see any problems.

cw's stdout is a terminal, when cw invokes gcc, gcc's stdout is a file

Ah, to deal with running several things?

or if -fdiagnostics-color= is passed in, it could pass that along and not add anything

yeah

which is why it's always disabled after bldenv

Yeah, I think probably: don't override if already passed, don't set it if NO_COLOR is set, and that's probably fine

i sort of figured that's why, but it was one of those things I always noticed (why aren't I geting color errors like elsewhere?) but never quite cared enough to really look until curiousity finally got me

[illumos-gate] 16979 stdbit.h _Generic pieces could work in C11 -- Robert Mustacchi <rm⊙fo>

jbk: one thing to watch out for is that make will itself redirect output to a per-job temporary file when building in parallel. Might need to peel another layer of that onion.

(but, yes, that sounds reasonable as long as it can be turned off -- I personally dislike colorized/angry fruit salad output)

rmustacc: I see a bit theory statement in usr/src/common/io/pciex/pcie.c (without looking, I"m going to guess it's from you :P) are there any other places I should look to start digging into those errors?

For digging into errors?

If that uts/common/io/pciex/pcie_fault.c.

"bit theory" I think you meant "big" (RM's theory statements aren't "bit". :) )

oh yeah, big.. the explanation there is really good and certainly not bit

Hello