Menu
▾
▴
#2476 fail with test148 apparently caused by bug introduced back in July, tandem datasets incl nominal
The Jenkins test148 calls file:/home/jbf/ct/autoplot/userlog/20170104/timeSeriesSpec.jyds?timerange=2017+Jan which has three aggregations which get the lat, lon, and ip from log files. I've found that starting with 20220629a, the ip comes back a shorter length. This shouldn't happen, and it's important that fill records be returned so several calls to tandem data sets need to return the same number of records.
Discussion
I don't know why this has suddenly started failing. The failure of the testing system needs to be studied as well.
I put in code to validate in the extent method, and this is now triggering an exception where an exceptional case was quietly ignored, before. (The problem is bins with min,max and there are points where the min is greater than the max.) I'll relax this check and allow them to go through for now. This is analogous to data marked as monotonic which is not actually monotonic, and I should see what validate does with that.