[gmx-developers] submitting to gerrit

Mon Jul 23 21:29:59 CEST 2018

Thank you for your responses. I did create a CR back in 2015, feature #1849, http://redmine.gromacs.org/issues/1849#change-12019. There wasn’t any feedback so I’m not sure if there is interest in it’s implementation. My code was working in late 2015 but I had a major set back, possibly due to bug #2264. I have completed and tested the project and could share my test cases through Jupiter Notebooks on Github. Publication is the next step. I have heard that there may be a change to expanded ensemble. My code works with the current version of expanded ensemble so I don’t know if there would need to be a revision in the near future. Your comments about sustainability/maintenance are understandable and I’m sure I would be able to triage and troubleshoot any related bugs as they come up. I would become more of an active member in the developer community as I’ve spent many hours understanding the code base but graduating is my current priority.

If I could get some feedback on feature #1849, then I will know whether or not it will be worthwhile for me push to get the code into the next code release.

Again, thank you for your responses.

-ChrisM

On Jul 18, 2018, at 11:40 PM, Erik Lindahl <erik.lindahl at gmail.com<mailto:erik.lindahl at gmail.com>> wrote:

And one more comment:

While we usually love improvements, fixes or a new algorithm, one important concern for all new code is how (and by whom) it will be maintained and bug-fixed a few years from now. A new analysis tool that doesn't change much other code can be quite straightforward, but a complex new algorithm that influences load balancing and the nonbonded kernels would need broad consensus from many developers that it's so useful we will support it together.

Since it can take a bit of effort to get documentation and unit tests ready, a good strategy is often to start a redmine feature thread about the algorithm and suggested contribution to get feedback earlier than after Gerrit submission!

Cheers,

Erik

Erik Lindahl <erik.lindahl at scilifelab.se<mailto:erik.lindahl at scilifelab.se>>
Professor of Biophysics
Science for Life Laboratory
Stockholm University & KTH
Office (SciLifeLab): +46 8 524 81567
Cell (Sweden): +46 73 4618050
Cell (US): +1 267 3078746

On 19 Jul 2018, at 06:24, Mark Abraham <mark.j.abraham at gmail.com<mailto:mark.j.abraham at gmail.com>> wrote:

Hi Chris,

On Tue, Jul 17, 2018 at 10:08 PM Mirabzadeh, Christopher (mira2978 at vandals.uidaho.edu<mailto:mira2978 at vandals.uidaho.edu>) <mira2978 at vandals.uidaho.edu<mailto:mira2978 at vandals.uidaho.edu>> wrote:
I have a completed project that was tested and has been working since GROMACS 5.1.4. I cloned the newest repo from gerrit and would like to try an include my code in the next code release. I understand that my code will have to fit the changes since 5.1.4.

Great. We love getting contributions of useful and well written code :-) Don't hesitate to drop into our fortnightly developer roundtable telcos. I'd offer to set up a specific chat with you, but I'm shortly going on my summer holidays! So perhaps others can carry on the discussion.

Question 1: Is it too late to include my code in the upcoming release? I remember getting an email from the group about the next release but got distracted by work and home and I may have missed the deadline.

No, it's not too late, but it very soon will be. Reviewable, tested code will need to be in gerrit by early September if we're to integrate it along with all the others things for GROMACS 2019. Do check out the advice in the latest version of the developer guide at http://jenkins.gromacs.org/job/Documentation_Nightly_master/javadoc/dev-manual/contribute.html. We definitely need to hear some more details from you about what you propose so we can see if it's feasible.

Question 2: I have no idea how to use gerrit. I’m familiar with git, but gerrit looks foreign to me. In the write-up on gromacs.org<http://gromacs.org/>, there is no mention of creating a local working branch and creating pull requests which is the workflow that I’m familiar with. Is that still the process with gerrit? Is there a workflow to follow that I missed? Further guidance would be appreciated.

We don't use a pull-request workflow. We haven't made a how-to for the mechanics of contributing, but all proposed patches (even from regular developers) are structured as a patch series based on the current HEAD of the master branch, and after testing and review get rebased on the new HEAD and get submitted. The history ends up linear. Working out how to organize the code so it isn't an unmanageable lump of 20k lines can be a bit of work. But we'd need to know more to have specific suggestions.

Question 3: If I have completed code for review, that incorporates a new algorithm, does a publication have to be included? The context being, what is the process for explaining the algorithm to a reviewer?

That will go a fair way to helping the code review process go smoothly, but do be aware that there will need to be other documentation for both users and developers, plus tests so that we can continue to keep the code working as we modernize the code base for future needs.

Mark

Cheers,

-ChrisM

--
Gromacs Developers mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers or send a mail to gmx-developers-request at gromacs.org<mailto:gmx-developers-request at gromacs.org>.
--
Gromacs Developers mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers or send a mail to gmx-developers-request at gromacs.org<mailto:gmx-developers-request at gromacs.org>.
--
Gromacs Developers mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers or send a mail to gmx-developers-request at gromacs.org<mailto:gmx-developers-request at gromacs.org>.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20180723/96066e78/attachment-0001.html>