Topics

Static Analysis And Windows CI Tests


Steve_G
 

Hi
For the last few days, 3 I think, everytime I update a branch in my fork it is running Static Analysis and  Windows Ci Tests.
This is very annoying and a waste of resouces, the local git, all branchs, are synced automatically every morning triggering at least 4 and sometimes 8 or more of these runs. Each run takes between 14mins to 6 hours, and including the master branch, which is just an exact copy of JMRI/master, normally fail.
Any ideas how I can stop this. 
Steve G.


Randall Wood <rhwood@...>
 

Update only what you working on and only when you working on it (i.e. only sync and push changes from the JMRI master into your fork only for the branch you are actively working on.)

Your other option is to not worry about it; you don’t pay for those resources, and these tests are not running on your computer.


On May 21, 2020, at 09:21, Steve_G <RailRodder@...> wrote:

Hi
For the last few days, 3 I think, everytime I update a branch in my fork it is running Static Analysis and  Windows Ci Tests.
This is very annoying and a waste of resouces, the local git, all branchs, are synced automatically every morning triggering at least 4 and sometimes 8 or more of these runs. Each run takes between 14mins to 6 hours, and including the master branch, which is just an exact copy of JMRI/master, normally fail.
Any ideas how I can stop this. 
Steve G.


Steve_G
 

Randall,
If I didnt keep my github uptodate and lose my workpc Im toast, and just because something appears free, as someone somewhere is paying for it, does not mean you can waste it.
Steve G.


Randall Wood <rhwood@...>
 

We can’t waste our time chasing down bugs that would have been caught by CI, so running the CI is less of waste of the JMRI contributor’s resources than not running would be.

If you want a CI-less JMRI we can do that, but be aware doing that will only turn away users and potential contributors.

Randall

On May 21, 2020, at 10:51, Steve_G <RailRodder@...> wrote:

Randall,
If I didnt keep my github uptodate and lose my workpc Im toast, and just because something appears free, as someone somewhere is paying for it, does not mean you can waste it.
Steve G.


Bob Jacobsen
 

I think you’re talking past each other.

Steve is asking if those CI pieces can be turned of then pushing individual commits (not a completed PR ready to submit) to his own repo copy (not the main JMRI/JMRI) one. He’s not asking to turn off CI on new PRs to master.

It’s certainly possible to believe that CI passing for a PR should be an absolute requirement before merging it; that’s not at all the same as requiring it for every commit.

Some people work so that every commit should pass CI, but many people don’t. I don’t, either in JMRI or my other projects.

Bob


On May 21, 2020, at 8:21 AM, Randall Wood via groups.io <rhwood=mac.com@groups.io> wrote:

We can’t waste our time chasing down bugs that would have been caught by CI, so running the CI is less of waste of the JMRI contributor’s resources than not running would be.

If you want a CI-less JMRI we can do that, but be aware doing that will only turn away users and potential contributors.

Randall

On May 21, 2020, at 10:51, Steve_G <RailRodder@...> wrote:

Randall,
If I didnt keep my github uptodate and lose my workpc Im toast, and just because something appears free, as someone somewhere is paying for it, does not mean you can waste it.
Steve G.
--
Bob Jacobsen
@BobJacobsen


Steve_G
 

On Thu, May 21, 2020 at 10:07 AM, Bob Jacobsen wrote:
Steve is asking if those CI pieces can be turned of then pushing individual commits (not a completed PR ready to submit) to his own repo copy (not the main JMRI/JMRI) one. He’s not asking to turn off CI on new PRs to master.
Yep Bob, thats all I wanted to know. Apparently its done in the settings for the fork, where is also lots of other interesting stuff. So now I can, like, travis, turn it on when Im getting close to something I may submit to merge, otherwise leave it off. 
Thanks
Steve G.


Bob Jacobsen
 

I went looking for that and didn’t find it. Could you provide a sentence or two on how to do that? I’ll put it on the (proposed) doc page. Thanks!

Bob

On May 21, 2020, at 10:31 AM, Steve_G <RailRodder@...> wrote:

Yep Bob, thats all I wanted to know. Apparently its done in the settings for the fork, where is also lots of other interesting stuff. So now I can, like, travis, turn it on when Im getting close to something I may submit to merge, otherwise leave it off.
--
Bob Jacobsen
@BobJacobsen


Randall Wood <rhwood@...>
 

The CI controls for GitHub are checked into the JMRI repo; if you fork the repo, you get the CI.

The only option i am aware of is to submit a PR to disable the CI for JMRI and getting approval to turn off CI for JMRI, or to ignore the fact you are getting CI run on your behalf.

Randall

On May 21, 2020, at 13:07, Bob Jacobsen <@BobJacobsen> wrote:

I think you’re talking past each other.

Steve is asking if those CI pieces can be turned of then pushing individual commits (not a completed PR ready to submit) to his own repo copy (not the main JMRI/JMRI) one. He’s not asking to turn off CI on new PRs to master.

It’s certainly possible to believe that CI passing for a PR should be an absolute requirement before merging it; that’s not at all the same as requiring it for every commit.

Some people work so that every commit should pass CI, but many people don’t. I don’t, either in JMRI or my other projects.

Bob


On May 21, 2020, at 8:21 AM, Randall Wood via groups.io <rhwood=mac.com@groups.io> wrote:

We can’t waste our time chasing down bugs that would have been caught by CI, so running the CI is less of waste of the JMRI contributor’s resources than not running would be.

If you want a CI-less JMRI we can do that, but be aware doing that will only turn away users and potential contributors.

Randall

On May 21, 2020, at 10:51, Steve_G <RailRodder@...> wrote:
Randall,
If I didnt keep my github uptodate and lose my workpc Im toast, and just because something appears free, as someone somewhere is paying for it, does not mean you can waste it.
Steve G.
--
Bob Jacobsen
@BobJacobsen






Bob Jacobsen
 

Found it!

I've added some instructions to the (proposed, draft) new CI page:

https://htmlpreview.github.io/?https://github.com/bobjacobsen/JMRI/blob/ci-text/help/en/html/doc/Technical/ContinuousIntegration.shtml

Please let me know what you think.

Bob


Steve_G
 

Yes, thats it. Looks fine. It might be worth while mentioning that if the "Actions" have been disabled then there is no "Actions"  tab between "Pull Requests" and "Objects".
Steve G.