Closed Ruhanga closed 1 year ago
@Ruhanga / @mks-d . This was already merged, so not sure you need or want my review still. No reason to oppose this, as long as it works and doesn't lead to performance issues, particularly at startup. I would like to know why this is necessary though - do we know what kind of situation works when we inject dependencies via autowiring and does not work when we explicitly inject using xml mappings? Seems odd.
@Ruhanga @mks-d I confirmed that this change (Autowired in a service) literally added 4 minutes to the start up time of our PIH EMR dev environments (haven't checked production instances but I assume the same as well).
This is due to this pesky issue that we've never figured out how to solve besides not autowiring in bean.
https://issues.openmrs.org/browse/TRUNK-3806
This is a blocker for our development, can we revert this until we figure it out, or at least revert and do a release before adding, because we need @mseaton 's drug CSV enhancement that only in this snapshot. Thanks!
Also fyi @ibacher
This is a followup PR to address a runtime bug caused by the InitializerDAO not being set. Issue: https://github.com/mekomsolutions/openmrs-module-initializer/issues/211