asset.opac_visible_copies is unused and should be removed

[Jeff Davis]

The asset.opac_visible_copies materialized view exists in current versions of Evergreen, but it's no longer used (I think it was superseded by opac.copy_vis_attr_cache and related developments in EG 3.0). On older Evergreen systems, asset.opac_visible_copies contains stale data; on newer systems, it will be empty. Either way, the view ought to be removed.

[Jeff Davis (jdavis-sitka)]

Note: asset.opac_visible_copies is used by the search.query_parser_fts() database function, which I believe is also obsolete and should be removed.

[Jeff Davis (jdavis-sitka)]

Working branch user/jeffdavis/lp1831803_remove_opac_visible_copies removes asset.opac_visible_copies and search.query_parser_fts (and some related Perl code that was already commented out).

https://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/jeffdavis/lp1831803_remove_opac_visible_copies

[Galen Charlton (gmc)]

I think this is on the right path. While reviewing it, I ran across a stored procedure, asset.refresh_opac_visible_copies_mat_view(), that updates asset.opac_visible_copies. It looks like there was never a schema update to remove that function, so that removal should be added to this patch series.

I think this also warrants a brief release notes entry in case there are any custom scripts floating out there that are (nowadays incorrectly) using this view.

[Jeff Davis (jdavis-sitka)]

Two additional commits have been pushed to the working branch: one to drop the asset.refresh_opac_visible_copies_mat_view() function if it exists, and another with release notes.

[Jason Boyer (jboyer)]

Hi Jeff, I tried this out and everything worked as expected and there were no complaints from postgres about missing funcs, signoff is here: https://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/jboyer/lp1831803_ovc_cleanup_signoff / working/user/jboyer/lp1831803_ovc_cleanup_signoff

The commit at the tip of that branch also removes the 2 data types that were apparently only ever used for search.query_parser_fts, and their removal also didn't cause any complaints.

[Galen Charlton (gmc)]

Pushed to master for inclusion in 3.9 along with a couple minor follow-ups. Thanks, Jeff and Jason!