On 04/29/2011 09:50 PM, James Mansion wrote:
Bit of an issue with the codepage for the single quotes here but it
should hold.
Someone has too much free time.
We know these and have fixed them (one of the first patches after 3.1.3
iirc).
server/server.cpp:223: error: ignoring return value of ‘ssize_t
write(int, const void*, size_t)’, declared with attribute
warn_unused_result
I feel like ranting today. I must have slept badly.
Adding that to -Wall is a disgrace. There's FAR too much code that
ignores the return values of write(), printf() and so on, and often
quite rightly (as in most of our cases — if we can't write debug stuff
to the console, what should we do? hold up the program until the console
is free? IMO forgetting about it is perfectly fine).
I think two people sent me "fixes" that just turned off -Wall.
Understandable. And NOT the kind of thing the compiler maintainers
should nudge people towards doing.
Seriously - waf is a LOT better than Jam (and a lot better than scons
too). Its very fast, and quite easy to extend, and having all of
Python's runtime available is very handy. Its a bit hard to hack on the
guts because its very 'dynamic pythonic' but Jam has a terrible
reputation for that sort of thing too, and Thomas Nagy is actively
developing waf.
Does it have jam's two great features: That it first decides what to
build, and only then runs commands, and that it builds in order of
source mtime, most recent first?
Arnt
