From: David Abrahams (gclbb-jamboost_at_[hidden])
Date: 2003-07-24 07:54:50
Vladimir Prus <ghost_at_[hidden]> writes:
> Uhm... I though something like that, but that did not explain strange
> behaviour on Boost.Book. Okay, I've wrote a test case:
> DEPENDS all : a ;
> DEPENDS a : b ;
> DEPENDS b : c ;
> TEMPORARY b ;
> actions copy
> cp $(>) $(<)
> copy a : b ;
> copy b : c ;
> and guess what. If "c" exists, then every "bjam" invocation recreates "a".
> I'm sure this is not right, or is there anything I've missed?
Maybe. If you look how it's used in Jambase, you get
libfoo.a -> libfoo.aX.o -> X.o -> X.c
-> libfoo.aY.o -> Y.o -> Y.c
-> libfoo.aZ.o -> Z.o -> Z.c
These don't actually exist, but are not marked NOTFILE (!). The .o
files are TEMPORARY. Hell if I know !?
You might try asking on the jamming list.
> This happens both with Perforce Jam and with Bjam (guess I've
> "merged" that bug from Perforce).
One of these days we're going to have to throw out Perforce Jam
-- Dave Abrahams Boost Consulting www.boost-consulting.com
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk