| Issue Number | 4989 |
|---|---|
| Summary | Build client files from schemas in GitHub branch |
| Created | 2021-06-10 09:19:43 |
| Issue Type | Improvement |
| Submitted By | Kline, Bob (NIH/NCI) [C] |
| Assigned To | Kline, Bob (NIH/NCI) [C] |
| Status | Closed |
| Resolved | 2021-06-24 11:07:38 |
| Resolution | Won't Fix |
| Path | /home/bkline/backups/jira/ocecdr/issue.291972 |
The current CDR build software generates DTDs and other client files from the schemas in the database for the current tier. That generally does what we want, because that tier is typically the development tier, on which the software about to be deployed to QA (and eventually STAGE and PROD) has been installed and unit tested. However, this is not completely reliable, and it is possible for a change to have been checked into the git branch but not installed on CDR DEV. It's also not possible to do a later build to verify that everything which is about to be deployed to STAGE or PROD matches what's in the GitHub branch, because at that point the DEV database is likely to have changes for future releases.
After reviewing the build/deploy software, I have come to the conclusion that
implementing the modification described above would require
extensive refactoring of the DTD/Schema parsing code in the
cdrapi/docs module (because the include and
import keywords are implemented with the assumption that
the included/imported schemas are in the database, not the file system);
and
it isn't necessary to make this change, because the deployment scripts first install the correct versions of all of the schemas in the CDR and then they rebuild all of the DTDs and document type information files.
Elapsed: 0:00:00.001551