[USRP-users] maint, master and auto clock rates

John Ackermann N8UR jra at febo.com
Mon Apr 20 16:36:26 EDT 2015


When using the build-gnuradio script (which I still find much more 
stable than pybombs, sorry), which branch is installed?

Thanks for your efforts managing all this!

John
----

On 4/20/2015 4:07 PM, Martin Braun via USRP-users wrote:
> As there were some recent confusions between our branches, I'd like to
> recap officially what our branches mean:
>
> * maint: This is the stable branch. All our releases come off of maint.
> In general, we only merge bugfixes into this branch, although we have
> had some exceptions to this rule (e.g. the UBX driver). API should never
> change on maint.
>
> * master: This is where new development happens. New devices get support
> on master first, and new features that change behaviour and/or API will
> go onto this branch. maint is merged into master, so consider this
> mostly a superset of maint -- except where master changes things.
>
> When we put out patch-level releases (e.g. 3.8.2 -> 3.8.3), we use the
> current state of maint, and tag that as the new release. master branch
> doesn't care about this.
>
> When we put out actual new versions (e.g. 3.8.x -> 3.9.0), things are
> different: maint gets reset to master, then we tag the version. So for
> every major release, maint and master are identical for a few weeks.
>
> One thing we recently changed is that master now gets a new version
> number. At this point, maint is identical to release 3.8.3, but master
> has version number 3.9.git (the .git showing that this is a development
> branch).
>
> On a side note, we might have other branches (e.g. the rfnoc-devel
> branch) that get other suffixes than .git.
>
>
> We recently had some B200-related changes that caused some confusion,
> which was related to this master/maint difference. In particular, on
> master, we changed these things:
> - The B200 has an auto-clock-rate feature, which will try to choose a
> clock rate based off of your chosen sampling rate.
> - The DSP chain changed on master, improving aliasing. We could have put
> this into maint, but it did also change group delay of the filter chain,
> hence we considered it too intrusive for maint.
>
> Cheers,
> Martin
>
> _______________________________________________
> USRP-users mailing list
> USRP-users at lists.ettus.com
> http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
>




More information about the USRP-users mailing list