J. Robinson wrote:
>
> Is something leaking?
Can't really say, because when you build off the tip, you're more likely
to get flakiness than if you build off a stable branch (e.g. 0.8.1).
Before pulling the code, it's a good idea to make sure the tree is
green; you should also check if there are any blockers that day which
you may have picked up; and you should see if any checkins occurred
between the time you started pulling the code and the time you finished.
Crossing paths with checkins can produce "interesting" effects (in the
may-you-have-an-interesting-life sense). I've found that while it's
morning on the West Coast and the tree is closed is generally the safest
(least hassles) time to pull; weekends too can sometimes be quieter in
terms of checkins. Now the new roadmap for landings/major checkins is
also a must-read before pulling.
>From the asserts you mentioned, I gather you've done a debug build.
Debugging code might also account for some of the shakiness. Full debug
emx builds became impossible last summer when layout got too big; they
only worked by disabling all the debug macros and just using -g, in
which case you don't get asserts and similar output. AFAIK, you're the
first person to do a full emx debug build since layout got split up, so
again you're into new territory.
If you're fairly certain you have a solid tree (i.e. no crossed paths
with checkins and no blockers), then you might try a nondebug build to
see if that gives you better results.
> Or am I doing the big bad by using pgcc instead of stock gcc?
Definitely not. [p]gcc 2.95.x is better for C++ than gcc 2.8.1, and
2.7.2 might even be a wiser choice than the latter because of the
bugginess generally associated with 2.8.x. IIRC, last time I did a 2.8.1
build was two years ago.
> Am I too full of silly questions?
If you are, I don't hear you asking them here.
h~
