Re: SIRF Release 0.9 Candidate

0 views
Skip to first unread message

Edoardo Pasca

unread,
May 8, 2017, 8:39:18 AM5/8/17
to CCP-PET...@jiscmail.ac.uk

Hi Kris,

 

On Friday we’ve made a few changes to the CCPPETMR_VM project on test_wm branch of the https://github.com/CCPPETMR/CCPPETMR_VM/tree/vm_test/scripts

 

I’ve built the VM using vagrant and the UPDATE.sh script.

 

Running the tests however doesn’t work because of the renaming we did of the environment variables. What is the plan now? Do I make the fixes to the SIRF code (pUtil.py, and corresponding Matlab files) in a new branch? I believe that the answer is Yes. Do I name it 0.9.0-rc2?

 

Edo

 

 

 

From: CCP-PETMR Developers list [mailto:CCP-PET...@JISCMAIL.AC.UK] On Behalf Of Thielemans, Kris
Sent: 05 May 2017 18:34
To: CCP-PET...@JISCMAIL.AC.UK
Subject: Re: SIRF Release 0.9 Candidate

 

Hi all

 

We’ve sent that message a tad too early…

 

I had forgotten to update our CCPPETMR fork of Gadgetron. I’ve done that now but unfortunately to a more recent commit than the one suggested by David (sorry. Mistake!). However, I hope this won’t be a problem as the SuperBuild was using most-recent Gadgetron already. I’ve now changed the SuperBuild to use the CCPPETMR Gadgetron fork (master).

 

I’ve now tagged SIRF and SIRF-SuperBuild repos to v0.9.0-rc1. 0.9.0-rc1 is therefore considered “frozen”.

 

As both the VM-updates and the Superbuild currently use the master branch of SIRF (and other forks), it means we should now refrain from committing to master (until our second 0.9 release candidate). New “good” commits on SIRF and SIRF-SuperBuild should be on a new branch v0.9.0-rc2 (commits under development can be on any branch obviously). Let’s discuss branching strategies on Thursday (not so easy with the superbuild&Travis etc!)

 

Thanks all for your help, especially Evgueni of course.

 

Kris

 

PS: If you check https://github.com/CCPPETMR/SIRF/releases, you’ll see that github automatically created a release for us (presumably because Casper was right to use v0.9.0-rc1!), zip and .tar.gz files of the source available.

 

From: Announcements and discussions for CCP PET/MR [mailto:CCP-...@JISCMAIL.AC.UK] On Behalf Of Evgueni Ovtchinnikov
Sent: 05 May 2017 15:13
To:
CCP-...@JISCMAIL.AC.UK
Subject: SIRF Release 0.9 Candidate

 

Dear All,

 

We are delighted to announce that SIRF Release 0.9 Candidate is now available on https://github.com/CCPPETMR.

 

Please visit http://www.ccppetmr.ac.uk/softwareframework.html, where you will find a link to Users Guide, the best place to start with our software, and further information.

 

At the moment, our installation scripts should work on Linux and Mac, installation on Windows is still very much work in progress.

 

Please try it out and let us know, by mailing to CCP-PET...@JISCMAIL.AC.UK, whether it worked for you and what is your impression.

 

We are now finalizing our new Virtual Machine, which will have all the necessary software (except Matlab) pre-installed and will have larger storage (64G). This work may affect the update process on the old VM, so please do not update it until further notice.

 

 

Thielemans, Kris

unread,
May 8, 2017, 7:20:11 PM5/8/17
to CCP-PET...@jiscmail.ac.uk

Hi

 

A bit confused about this. I didn’t think those renaming should affect the code. I would certainly not change pUtil etc. Potentially we have to rename the variables back. However, I suspect this was because we forgot to put SIRF_PATH in the new UPDATE.sh (which I discovered later and fixed already https://github.com/CCPPETMR/CCPPETMR_VM/commit/d212311b0a0356ccccc3ae81c9a70f928653687a)

If not, can you give some more info?

 

Kris

Edoardo Pasca

unread,
May 9, 2017, 4:35:00 AM5/9/17
to CCP-PET...@jiscmail.ac.uk

The problem is here

https://github.com/CCPPETMR/SIRF/blob/master/src/common/pUtil.py#L25

 

neither of the two (SIRF_PATH, SRC_PATH) environment variables are defined now.

 

You are right and the fix you mention should do the trick.

 

Edo

Reply all
Reply to author
Forward
0 new messages