Subject: Re: [boost] [Boost.Sync] no master branch?
From: Eric Niebler (eniebler_at_[hidden])
Date: 2014-01-15 14:21:50
On 01/15/2014 10:57 AM, Tim Blechmann wrote:
>>> I discovered this anomaly by trying to check out
>>> every repo in the master branch with this git command:
>>>> git submodule foreach git checkout master
>>> That command fails today because of sync.
>> It shouldn't, sync isn't a submodule of the master branch. Try using
>> 'git submodule update --init'.
> not sure if this will work: especially a user could have modifications
> to the submodule, but re-initializing all submodules to the state that
> the supermodule expects will require git to remove the submodule. i'm
> pretty sure that git will not like (read reject) this.
> in my personal experience, git submodules are references that one rarely
> wants to change and never wants to remove from the working copy, though
> the situation may have improved these days.
> an empty master branch may be a workaround, though.
This is probably now a discussion about best practices wrt testing
modularized boost, but I want to test my libraries against what will
comprise the next official boost release. To me, that suggests checking
out develop on the super-project and master on all sub-projects. Have I
got that all wrong? I admit to being a bit of a n00b here.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk