[Rivet] Rivet release plan

[Andy Buckley]

On 07/05/14 11:28, Frank Siegert wrote:
> Hi Andy, all,
> 
>> Given the embarrassing issue with the abspid() function that wasn't abs,
>> as well as the CMS fixes, another CMS analysis to add, a fix to some
>> HepData-produced ref data, and a few other tweaks... I think it's about
>> time to release another Rivet update, 2.1.2. I'll try to get the
>> necessary things done for that in the next week.
> 
> I've played around a bit with Rivet/YODA and Python3 last week (due to
> an updated OS installation). It was a bit of a struggle to get it
> working, and I'm not 100% finished yet (e.g. with some of the scripts
> like yodamerge), but at least the build system and pyext stuff seems
> to be working.
> 
> What are our Python3 compatibility plans... as early as possible in a
> bugfix release, or rather aiming for the next larger release (e.g.
> 2.2.0) to make that available? I think some of the changes need to be
> tested well on different systems and need more Python2/3 compatibility
> wrapping than I have put in for my testing so far.

David asked about this last week, and I have to admit that I was
ignoring the issue on the assumption (correct as far as I know) that the
LHC experiments are not making any moves at all toward Python3. But
obviously it would be nice if we can be compatible.

If we are already requiring Python 2.5 then I think there are "from
future import ..." mechanisms to pull in the print function, division
behaviour, and maybe some of the changed types.

The big 2->3 switch is usually between strings and bytes, from what I've
heard: I hope that's largely irrelevant to us, but can't be sure without
trying: what were the issues that you have to deal with?

> As a short term measure, I wonder whether we should follow PEP 394
> (http://legacy.python.org/dev/peps/pep-0394/) and simply call python2
> explicitly in all scripts:
> 
>   "- however, end users should be aware that python refers to python3
> on at least Arch Linux (that change is what prompted the creation of
> this PEP), so python should be used in the shebang line only for
> scripts that are source compatible with both Python 2 and 3
> - in preparation for an eventual change in the default version of
> Python, Python 2 only scripts should either be updated to be source
> compatible with Python 3 or else to use python2 in the shebang line."
> 
> I have checked that python2 is actually available in SLC5 and SLC6,
> but I don't know whether we find any platform where that's not the
> case (OS X?).

It's also there on Ubuntu 13.10... not sure when it first appeared, but
it was missing (ridiculously) a couple of years ago so maybe the LTS
edition that's installed on some HEP sites does not have a python2
symlink? David, what's the situation on the IPPP desktops?

OS X people, with and without HomeBrew/MacPorts/etc., do you have a
python2 executable?

If all these systems have python2 then I think we should switch our
scripts to use that in the shebang line for 2.1.2.

>> Lastly, I'll just mention that we will have a new postdoc starting in
>> Glasgow in August, a substantial part of whose time will be working on
>> Rivet. So maybe we can organise a "sprint" shortly after he arrives.
>> Having one before that would also be good, though... anyone want to host?
> 
> If you don't mind the suboptimal flight options, I'd offer to host one
> of the next ones in Dresden, but only after October, as I might even
> have some money then to cover at least some of the local costs.
> As for one in/around August, I'm not available July 26 - Aug 6/9 (if
> we happen to do it in Kopenhagen/Lund(?), Aug 6-8 would work great for
> me, as I have to be in that area anyway around that time). After that,
> and especially also in September, it looks quite good.

I have (or rather my wife has...) some weddings to go to in southern
England in the first week of August, so I was planning to maybe go to
UCL for a couple of days. Not sure if popping over to Lund will be an
option, but maybe...!

I'd be happy to host something in Glasgow in August/September, and maybe
we aim for a Dresden-based sprint later in the year/early 2015?

Andy

-- 
Dr Andy Buckley, Royal Society University Research Fellow
Particle Physics Expt Group, University of Glasgow / PH Dept, CERN