Subject: Re: [Boost-build] [Boost-commit] svn:boost r79780 - trunk/tools/build/v2/engine
From: Jurko GospodnetiÄ (jurko.gospodnetic_at_[hidden])
Date: 2012-08-02 04:12:49
>> Updated Boost Jam to know how to report its minimum supported
>> file modification timestamp resolution (currently reported as
>> part of Boost Jam's version information). This allows external
>> tools using Boost Jam to adapt to Boost Jam's potential ignorance
>> of fine file modification timestamp changes.
> I don't think the version output is the correct
> place to put this. Please find a better way.
Do you have any suggestions?
I see the '-v' option as asking Boost Jam to 'tell me things about
itself that might be important to me as a user'. And this information is
important to me 'as a user' as it tells me how long I have to wait to
have the Build system recognize a touched/modified file as such.
I looked into adding this as a separate command-line option but there
is already a mess of those, the existing ones did not seem to fit the
purpose other than '-v' and there were no unused letters available that
seemed to fit either.
YMMV, but at the moment, I really do not see a place where this
information would be better presented. '-v' option reports Boost Jam's
version information including the version identifier, the OS and some
arbitrary list of copyright holders. Whether a specific implementation
supports finer grained path modification timestamp resolution seems like
another implementation detail information token at the same level as those.
Confusingly, we also have a different '--version' option handled by
Boost Build code and reporting a bit different information, but I see
that as a separate issue.
>> Boost Jam version information output cleaned up a bit to make it
>> easier to update it with additional information.
> /Really/? If you want some kind of key/value
> feature testing, it doesn't belong under the
> regular version flag. Please revert this.
> I just find it ugly.
If you do not like the formatting - please suggest a better one. Or,
if you /Really/ feel that strong about the current one and can give no
alternate suggestions, I can just revert the whole shabang, but please
first confirm explicitly whether this is needed as I'm not really happy
with that course of action...
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