Boost logo

Boost-Build :

From: David Abrahams (dave_at_[hidden])
Date: 2004-11-16 11:40:25


"c_andy_martin" <c_andy_martin_at_[hidden]> writes:

> OK. I've done that, and as I suspected, somehow the source files end
> up with dependencies on the directory common/exception. I suspect it
> has to do with the dependency scan somehow picking up #include
> <exception> as pointing to the directory 'exception'. I will try to
> include some of the relevant output (this is using the project
> heirarchy from my original message which has a different subject line).
>
> Output from bjam -d+12 (I have omitted portions with ...). This output
> was taken after I had run bjam several times (so that no targets were
> being updated). I then ran `touch common/exception' followed by `bjam
> -d+12'.

<snip>

> And, the relevant jamfiles:
> What first tipped me off to this behavior was running `vi' in the
> exception directory always triggered a complete rebuild. This is
> becuase `vi' deposited a .swp file in the exception directory, which
> causes the directory's timestamp to be newer than the source,
> triggereing a rebuild. That's why the project is in the directory
> `vi-boost-interaction' because I was trying to figure out why vi would
> cause things to rebuild.
>
> Thanks for your help!!!
> -Andy

Sorry, I just wasn't thinking. The collision here is between the
header <exception> and the first path component of
<exception/Exception.h>. Okay, that is fortunately easy to fix. All
we need to do is fix the rules that make directory targets so that
those targets are gristed slightly differently.

Hmm...

Please try this, with v1. Just add

-sdirectory-grist=FuBar

to your bjam command-line. Do you still see these problems?

-- 
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