|
Boost-Build : |
Subject: Re: [Boost-build] Generating multiple CPP/H targets from a single source
From: Fabien Chêne (fabien.chene_at_[hidden])
Date: 2010-08-16 06:08:05
2010/8/14 Vladimir Prus <ghost_at_[hidden]>:
> On Friday 13 August 2010 13:13:59 Fabien Chêne wrote:
>
>> [...]
>>
>> The bug is quite complex to describe, and I can imagine that I am not
>> perfectly clear.
>> So, I have created a reduced testcase that triggers the problem.
>>
>> Any help would be greatly appreciated.
>
> Thank you for a testcase. Could you please try the attached patch?
> It fixes your testcase, and if it fixes your real project as well,
> I'll commit it.
Many thanks for this patch. It fixes the testcase for sure. About 'my'
real project, it now succeeds in all compile test I have written
(bjam -a, removing only .h, only .hh, only .h and .hh, only .cxx, only
SK.cxx, re-build aftier clean, etc) at any level of parallelism
(-j1,2,4,8,16,32).
It previously fails for half of all tests. So, it is a slight improvement.
Unfortunately, I run into another problem with this patch. If bjam is
invoked from the toplevel of my project, it fails bawling something
like: 'actions too long (more than 102400)'. Indeed there are a lot of
sources in this library, with very long path. Apparently, something --
perhaps not relative to the correctness of the patch -- causes the
command line to increase a bit. Is it possible to increase MAXLINE ?
(With MAXLINE in jam.h set to 112400, it works for me).
-- Fabien
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