From: David Abrahams (dave_at_[hidden])
Date: 2004-07-20 07:28:30
Daniel Schlyder <daniel_at_[hidden]> writes:
> David Abrahams:
>> ...when we were discussing the merits/disadvantages of rewriting bjam
>> (perhaps on top of Scons), I forgot this one: people continue to
>> occasionally report bjam crashes, and we have *no clue* as to what
>> code is causing the problems.
> Reading this made me remember a while ago I started having problems with bjam
> (v1) crashes, which seemed to have been triggered by the number of targets
> growing considerable at that time. I tried both pre-built binaries and building
> from source without luck. It wouldn't crash all the time; sometimes I could run
> it half a dozen times before it happened. I finally figured out it had
> something to do with bin\.jamdeps. I wrote a wrapper script that deleted this
> file before running bjam and have had no crash since.
Hum. Well, that's an embarrassment to me, since I was the one who
ported that patch over from Craig McPheeters' fork of Perforce Jam.
It's supposed to speed up builds by cacheing dependency information,
(but of course if you're deleting the file it won't have any effect).
However, I never vetted the code for correctness and I haven't tried
to find any updates from its author since May 2002. Craig, any news?
-- Dave Abrahams Boost Consulting http://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