- Koha-bugs - lists.koha-community.org

[Bug 43041] Uncaught exceptions and DBI errors are not timestamped in plack-error.log
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=43041 --- Comment #4 from Martin Renvoize (ashimema) <martin.renvoize(a)openfifth.co.uk> --- Created attachment 201863 --> https://bugs.koha-community.org/bugzilla3/attachment.cgi?id=201863&action=e… Bug 43041: Uncaught exceptions and DBI errors are not timestamped in plack-error.log Bug 16357 added timestamps to Plack warnings by sending them through Log4perl with the LogWarn middleware. Only warnings take that path. Uncaught exceptions, DBI errors, and anything a script prints to STDERR ( technically nothering does that, but reports/catalogue_stats.pl has some debug code that could in theory ) are written to the psgi.errors stream, so they end up in plack-error.log with no timestamp. This makes it hard to tell when an error happened, or to find all the errors for a given day. This patch adds a tied psgi.errors filehandle that sends those writes through Log4perl, so they are logged with a timestamp in plack-<interface>-error.log alongside the warnings. Test Plan: 1) Add a line that dies to the top of a staff script such as mainpage.pl: die "Bug XXXXX test"; 2) Restart plack 3) Load the staff home page 4) Note the exception is written to plack-error.log with no timestamp 6) Apply this patch 7) Restart plack again 8) Reload the staff home page 9) Note the exception is now written to plack-intranet-error.log with a timestamp, and no longer written to plack-error.log! 10) Change the die to a warn 11) Retart plack again 12) Reload the staff home page 13) Verify the warns still come through as they did pre-patch Signed-off-by: Martin Renvoize <martin.renvoize(a)openfifth.co.uk> -- You are receiving this mail because: You are watching all bug changes.

1 0

[Bug 43041] Uncaught exceptions and DBI errors are not timestamped in plack-error.log
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

1 0

[Bug 43041] Uncaught exceptions and DBI errors are not timestamped in plack-error.log
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=43041 --- Comment #5 from Martin Renvoize (ashimema) <martin.renvoize(a)openfifth.co.uk> --- Created attachment 201864 --> https://bugs.koha-community.org/bugzilla3/attachment.cgi?id=201864&action=e… Bug 43041: Make intranet_svc mount catch and log uncaught exceptions For reasons unkown, /intranet_svc doesn't have the HTTPExceptions middleware enabled so an uncaught exception in a service script ends up in plack-error.log Test Plan: 1) Add a line that dies to the top of a service script such as svc/checkouts: die "This is a test"; 2) Restart plack 3) Call the service, e.g. load the checkouts tab for a patron, which calls /cgi-bin/koha/svc/checkouts 4) Note the exception is written to plack-error.log with no timestamp 5) Apply this patch 6) Restart plack again 7) Repeat the call 8) Note the exception is now written to plack-intranet-error.log with a timestamp! Signed-off-by: Martin Renvoize <martin.renvoize(a)openfifth.co.uk> -- You are receiving this mail because: You are watching all bug changes.

1 0

[Bug 43041] Uncaught exceptions and DBI errors are not timestamped in plack-error.log
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

1 0

[Bug 8589] Advanced Search for itemtype and location produces some results that are not useful
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=8589 Luis Bataller <luis.bataller(a)xercode.es> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |luis.bataller(a)xercode.es --- Comment #9 from Luis Bataller <luis.bataller(a)xercode.es> --- Hi all, This bug captures a limitation of searches in Koha under Elasticsearch: filtering by item-level attributes such as itemtype and location returns records where each criterion is satisfied by a different item, rather than by the same physical item. The same cross-item mismatch affects branch/availability filters. We have implemented a solution against current Koha (26.05) and would like to offer it to the community. This is essentially the approach David Cook anticipated in comment #7: Elasticsearch's nested data structures let you require an "AND" across values within the same nested object. We index each record's items as a nested object on the biblios index (alongside the existing flat 952 fields, which are kept untouched), so that item-level conditions are evaluated on the same item. What the work covers so far: Nested item mapping on the biblios index, populated with per-item fields (home/holding branch, itype, location, ccode, callnumber, barcode, acquisition date, lost/withdrawn/damaged, plus derived flags: availability, in-transit, on-hold, on-request, and home/holding geolocation as geo_point). Same-item filtering for facets, branch/multibranchlimit limits and availability — the fix for this bug. Selecting itemtype + location now returns only records with an item matching both. Advanced-search item operands (barcode, location, call number, acquisition-date range) and the item-type limit routed through the same nested channel when the query is a pure-AND (documented fallback to flat for OR/NOT). Reactive reindexing on transfers, holds and library-geolocation changes, so the derived fields stay in sync. OPAC item hiding and record suppression honoured through the nested channel (OpacHiddenItems, hidelostitems, OpacHiddenItemsHidesRecord), including a plugin hook for custom hiding rules. Optionally, showing only the matching items in staff search results (with recalculated counts), behind a new syspref. Together these cover the case Ray Delahunty raised in comment #8: the keyword area, the tab operands and the bottom limits of Advanced Search working together without "not useful" results, including complex multi-site setups where facet (homebranch) filtering currently still leaks cross-item hits. We would be happy to attach the patch series here. Given the size of the change, before uploading everything, we'd appreciate the maintainers' guidance on how you'd prefer to receive it: as a single dependency chain, split into different staged bugs (mapping, filtering, facets …), or whether an RFC is warranted first. Happy to adapt to whatever makes review easiest. Thanks! -- You are receiving this mail because: You are watching all bug changes.

1 0

[Bug 43041] Uncaught exceptions and DBI errors are not timestamped in plack-error.log
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=43041 Kyle M Hall (khall) <kyle(a)bywatersolutions.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |Needs Signoff -- You are receiving this mail because: You are watching all bug changes.

1 0

[Bug 43041] Uncaught exceptions and DBI errors are not timestamped in plack-error.log
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=43041 --- Comment #2 from Kyle M Hall (khall) <kyle(a)bywatersolutions.com> --- Created attachment 201819 --> https://bugs.koha-community.org/bugzilla3/attachment.cgi?id=201819&action=e… Bug 43041: Make intranet_svc mount catch and log uncaught exceptions For reasons unkown, /intranet_svc doesn't have the HTTPExceptions middleware enabled so an uncaught exception in a service script ends up in plack-error.log Test Plan: 1) Add a line that dies to the top of a service script such as svc/checkouts: die "This is a test"; 2) Restart plack 3) Call the service, e.g. load the checkouts tab for a patron, which calls /cgi-bin/koha/svc/checkouts 4) Note the exception is written to plack-error.log with no timestamp 5) Apply this patch 6) Restart plack again 7) Repeat the call 8) Note the exception is now written to plack-intranet-error.log with a timestamp! -- You are receiving this mail because: You are watching all bug changes.

1 0

[Bug 43041] Uncaught exceptions and DBI errors are not timestamped in plack-error.log
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=43041 --- Comment #1 from Kyle M Hall (khall) <kyle(a)bywatersolutions.com> --- Created attachment 201818 --> https://bugs.koha-community.org/bugzilla3/attachment.cgi?id=201818&action=e… Bug 43041: Uncaught exceptions and DBI errors are not timestamped in plack-error.log Bug 16357 added timestamps to Plack warnings by sending them through Log4perl with the LogWarn middleware. Only warnings take that path. Uncaught exceptions, DBI errors, and anything a script prints to STDERR ( technically nothering does that, but reports/catalogue_stats.pl has some debug code that could in theory ) are written to the psgi.errors stream, so they end up in plack-error.log with no timestamp. This makes it hard to tell when an error happened, or to find all the errors for a given day. This patch adds a tied psgi.errors filehandle that sends those writes through Log4perl, so they are logged with a timestamp in plack-<interface>-error.log alongside the warnings. Test Plan: 1) Add a line that dies to the top of a staff script such as mainpage.pl: die "Bug XXXXX test"; 2) Restart plack 3) Load the staff home page 4) Note the exception is written to plack-error.log with no timestamp 6) Apply this patch 7) Restart plack again 8) Reload the staff home page 9) Note the exception is now written to plack-intranet-error.log with a timestamp, and no longer written to plack-error.log! 10) Change the die to a warn 11) Retart plack again 12) Reload the staff home page 13) Verify the warns still come through as they did pre-patch -- You are receiving this mail because: You are watching all bug changes.

1 0

[Bug 40619] New: Remove OverDrive star ratings from OPAC JS
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=40619 Bug ID: 40619 Summary: Remove OverDrive star ratings from OPAC JS Change sponsored?: --- Product: Koha Version: Main Hardware: All OS: All Status: NEW Severity: minor Priority: P5 - low Component: OPAC Assignee: oleonard(a)myacpl.org Reporter: oleonard(a)myacpl.org QA Contact: testopia(a)bugs.koha-community.org OverDrive no longer supplies star ratings via their API: https://developer.overdrive.com/blog/16-may-2025-coming-soon-adding-content… We can remove the code in opac-overdrive-search.tt which used to show those star ratings. -- You are receiving this mail because: You are watching all bug changes.

1 15

[Bug 18139] New: 'Too many checked out' can confuse librarians
by bugzilla-daemon＠bugs.koha-community.org 12 Jul '26

12 Jul '26

https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=18139 Bug ID: 18139 Summary: 'Too many checked out' can confuse librarians Change sponsored?: --- Product: Koha Version: unspecified Hardware: All OS: All Status: NEW Severity: minor Priority: P5 - low Component: Circulation Assignee: koha-bugs(a)lists.koha-community.org Reporter: kyle(a)bywatersolutions.com QA Contact: testopia(a)bugs.koha-community.org CC: gmcharlt(a)gmail.com, kyle.m.hall(a)gmail.com The too many checked out message can be very confusing for librarians when the limit specified is more or less than the number of checkouts the patron already has. We should make it clear to the librarian that this is only counting certain types of checkouts by giving them the issuing rule criteria used. -- You are receiving this mail because: You are the assignee for the bug. You are watching all bug changes.

1 54