From: Beman Dawes (bdawes_at_[hidden])
Date: 2002-07-25 08:48:50
At 07:24 AM 7/25/2002, David Abrahams wrote:
>From: "Beman Dawes" <bdawes_at_[hidden]>
>> One of the challenges remaining for bjam-based regression testing is
>> able to detect that a test has failed because a library it depends on
>> failed to build, and to report the compiler or other error messages
>> indicate why the library build failed.
>> The current system already provides the error messages. What is missing
>> a notification like:
>> "x failed because it depends on y, which failed".
>> x and y are the target directory paths.
>The system already says "skipping <x> due to missing <y>"
>The paths are not /exactly/ represented, but you can figure them out from
><x> and <y>.
Ah! I'd missed those messages in the clutter.
It will take a bit of deduction to eliminate the "skipped" messages
unrelated to library builds, but that isn't a problem for now.
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