Subject: Re: [Boost-build] Early bjam abort
From: Steven Watanabe (watanabesj_at_[hidden])
Date: 2012-05-01 11:52:20
On 05/01/2012 08:28 AM, Nogradi, Chris wrote:
> On Tuesday, May 01, 2012 1:46 AM, Steven Watanabe wrote:
>> Okay. Try r78287. I've been running
>> randomly generated configurations for
>> the last half-hour without any errors
>> with this version.
> Thanks for fixing these issues so quickly. I have tried this change and it behaves reasonably well. It no longer aborts prematurely (at least I have not seen an instance of it yet) and it seems to properly handle the dependency order. I do have a couple of issues that yet remain. I have one instance of a generated C file that depends on a generated header file and for some reason the dependency does not work. I am still investigating this issue since it was present before your changes and may just be something I am doing wrong. The second issue is that when I do a build immediately after a full build has completed (i.e. w/o any additional changes), I would expect everything being up to date. Instead, there are a hand full of C files (always the same ones and not generated ones) that are always rebuilt. This behavior is consistent across all my builds and never converges. I will try to come up with a scaled down example of this for you.
-d+12 will show exactly why each target is updated.
Also, check that you don't have any files
with a timestamp in the future.
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