search

spacetelescope / drizzlepac

AstroDrizzle for HST images.
https://drizzlepac.readthedocs.io
BSD 3-Clause "New" or "Revised" License
50 stars 39 forks source link

Build: Test patch for S_REGION is HSTDP-2024.3.1 #1885

Open stscijgbot-hstdp opened 3 weeks ago

stscijgbot-hstdp commented 3 weeks ago

Issue HLA-1336 was created on JIRA by Michele De La Pena:

Work with Steve (who is the lead) and Joe, to generate a Quick Fix build which only will contain the fix to the S_REGION keyword.  This build should not be allowed to pick up new versions of packages (domestic and third-party), if at all possible, so that INS does not need to be involved with regression testing.  This build will only have an update to Drizzlepac 3.7.1.1 for the S_REGION fix.

Since we will be using STASIS (a new build system) in a restrictive way (hold domestic and third-party packages to the versions in HSTDP-2024.2.0), it is anticipated we will encounter some issues.

Note the build name from SCSB is: CALDP-2024.10.1-1+mostlyCOS-py311-linux-x86_64.yml release_fmt = %n-%v-%r+%c-py%p-%o-%a Formatter Purpose n Delivery name (CALDP) c Delivery codename (mostlyCOS) r Delivery RC level (1) R Delivery RC level (final-aware) v Delivery version (2024.10.1) m Mission name P Python version (i.e. 3.11) p Compact Python version (i.e. 311) a System CPU Architecture (x86_64) o System Platform (OS) (linux) t UNIX Epoch timestamp

stscijgbot-hstdp commented 1 week ago

Comment by Michele De La Pena on JIRA:

Joe and I discussed generating the S_REGION patch with STASIS even though he knows there are some features which do not work. These known features are "extras" and are not necessary to generate this basic build. He tried generating the build YML. The processing completed successfully, but upon checking the YML, the process did not (1) pull in the correct version of drizzlepac, and (2) it went to GitHub repo for a package (cannot remember which one) instead of getting it from PyPi or Condaforge.

I suggested he stand down for the weekend and re-group so he can attack the issue on Monday with a clear head.

stscijgbot-hstdp commented 1 week ago

Comment by Michele De La Pena on JIRA:

Joe was able to use STASIS to generate the build for the Linux system. I examined the YML file and found differences (additional packages), but I was looking at the wrong YML. I was looking at the "test" YML which contains the entire test package manifest that was used to test the environment, that includes pytest and everything else that was required at the time. This is not in the actual build YML that DSB or I will use. The only pip package differences were the drizzlepac (3.7.1 - 3.7.1.1) and the new package, simplify_polyline. There were library differences, < - libstdcxx-ng=14.1.0=h4852527_1

  • libstdcxx-ng=14.1.0=hc0a3c3a_0 but these should not matter. I created an environment on Linux and tested that it works for calacs.e. Steve is going to test the S_REGION change.