Reference build on Jenkins

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Reference build on Jenkins

Erick Erickson
Several issues:

1> I think Hoss’ rollups pull in those errors. In that case, the BadApple report is essentially useless. I’ll keep sending it out since it only takes a few minutes to create and send. I’m not complaining about this, I expected it. I think it’s worthwhile to keep sending out to monitor progress. If we’re _really_ lucky, we won’t need it any more in a month or two…

2> The source pattern and warning checks fail.

2a> The ref impl was cut in June IIRC, when I was only partway through the suppresswarnings. I’m not particularly looking forward to doing some subset of 8,000 warnings over again, it’s mind-numbing. Fortunately I have macros. If I’m going to do any of that again I want a guarantee that the work won’t have to be repeated yet again. Is the reference_impl_dev the one that’ll be used going forward? If changes are made there, are they guaranteed not to be lost?

2b> Or will many/most of these warnings go away as the branches are merged? In which case I should just cool my jets until the merge is complete and worry about it then.

2c> If we’re going to get any mileage out of Jenkins near-term, we need to turn off those checks. I’ll veto any suggestion that they be disabled _permanently_, but while the tests shake out and the branches get merged, I think they should be disabled. Progress on the test stability is more important than these checks succeeding for the time being.



---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]