From: Eoin (eoin-keyword-boost.jam.dca6c0_at_[hidden])
Date: 2006-08-18 12:31:53
Reece Dunn wrote:
> Eoin wrote:
>> Hello, I recently started using the latest CVS version of Boost to test
>> my most recent project and ensure it's compatible, in the process of
>> doing so I've gotten the impression that the latest CVS version of
>> Boost.Build v2 is slower to get going over the one in 1.33.1;
>> specifically there is a longer pause between when you execute bjam and
>> when it prints "warning: Python location is not configured" and the
>> other initial messages.
> This is a known issue.
>> Just thought I'd share this to see if anyone else has noticed the same
>> behavior or perhaps instead I've done something wrong with my setup.
> It is not your setup. When invoking the top-level Jamfile.v2 script in the
> Boost directory, BBv2 is globbing for header files to setup a header target,
> allowing you to install the headers through BBv2. This is one of the things
> that is causing the slowing. There was a thread on this a short while back.
> There have also been people who are using a different hashing algorithm
> to the one that bjam currently uses which is reported to improve the
> Other things can be used to speed up BBv2, but ultimately, the bjam
> language isn't designed for what BBv2 is doing - BBv2 is pushing the
> bjam language to its limits. This is one of the motivations for moving
> Boost.Build to Python.
Thanks for the response, while its a pity that things are slower it is
promising that an alternative is being investigated, I look forward to
the future of BBv2.
Kind regards, Eoin.
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