This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

full regression test size

With the recent addition of new flag combinations to make check,
the size of it is starting to be a problem...

- Between make bootstrap that leaves stage1, stage2 files lying around,
- between all the cores that the various aborts leave around,

it becomes really trying... the latest snapshot went up to >200Mb on my
i386. I shudder at what a risc architecture without shared libraries will
come up to.

I think that the cores could probably be wiped out while making check
(after all, any core that needs investigating is pretty trivial to 
reproduce) and there could possibly be another target than make bootstrap
that wipes stage1/* stage2/* after building (pretty simple, maybe I've
missed it)... I'm not talking about bootstrap-lean, since I want to have
the intermediate files in case there's a problem.

This may help get more results... right now, there's a sizable set of
people who just try out snapshots and report problems without knowing how
to fix them; with the current size, this is getting more and more ackward.

	Marc Espie		
|anime, sf, juggling, unicycle, acrobatics, comics...
|AmigaOS, OpenBSD, C++, perl, Icon, PostScript...
| `real programmers don't die, they just get out of beta'

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]