|
Boost-Build : |
From: David Abrahams (dave_at_[hidden])
Date: 2007-05-24 10:11:44
on Thu May 24 2007, Vladimir Prus <ghost-AT-cs.msu.su> wrote:
> On Wednesday 16 May 2007 21:25, David Abrahams wrote:
>>
>> on Tue May 15 2007, Vladimir Prus <ghost-AT-cs.msu.su> wrote:
>>
>> > Over at IRC, and on mailing list, it seems --toolset option confuses
>> > folks -- they extrapolate the syntax and try to use "--link=static"
>> > and so on.
>> >
>> > I think we probably better either:
>> >
>> > 1. Rename --toolset to --autoconfigure-toolset, or something.
>> > It might be better to disable auto-configuration for "toolset=foo" syntax
>> > at the same time, to make "toolset" feature no longer "special".
>> >
>> > 2. Remove "--toolset" completely. Retain autoconfiguration
>> > for "toolset=foo" and tell users to use "toolset=foo".
>> >
>> > For reasons I don't understand, I prefer (2). Comments?
>>
>> I don't have a strong opinion, but I think the presence of
>>
>> foo=xxx
>>
>> and
>>
>> --bar=baz
>>
>> or even just
>>
>> --bar
>>
>> on the same command-line will be confusing. New users don't have a
>> conceptual grasp of the difference between features and command-line
>> options, and even more experienced users can't keep track of all the
>> feature names.
>
> On the other hand, there *is* fundamental difference between giving
> a value of a feature, and just random option. Features have some common
> semantics and they appears in target path, while option can have any effect
> whatsoever.
>
> It should be noted that we haven't had this confusing before, when
> --toolset *option* was not introduced, which makes me thing it's
> --toolset that confuses people.
I'm not opposed to changing that, but for the record, I still think
there is a deeper conceptual problem on the users' part that we're
encouraging.
-- 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