- Add a file-level comment reminding that every shipped output.path needs a
matching 'NONE' target-map entry and cp step in Makefile.PL
Signed-off-by: Paul Derscheid <paul.derscheid@lmscloud.de>
Signed-off-by: David Cook <dcook@prosentient.com.au>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
The OPAC vue bundle (koha-tmpl/opac-tmpl/bootstrap/js/vue/dist) was not
handled like its intranet counterpart in Makefile.PL: exclude it from the
generic target map and copy it explicitly into blib in move_compiled_js.
Test plan:
1. In a KTD shell, from the koha checkout:
perl Makefile.PL
(accept the defaults, 'standard' install mode is fine)
2. make -n move_compiled_js | grep 'js/vue/dist'
=> without this patch only the intranet cp line is printed;
with it there is also a line copying
koha-tmpl/opac-tmpl/bootstrap/js/vue/dist into
blib/OPAC_TMPL_DIR/bootstrap/js/vue/
3. make
4. ls blib/INTRANET_TMPL_DIR/prog/js/vue/dist blib/OPAC_TMPL_DIR/bootstrap/js/vue/dist
=> both directories exist and contain the compiled bundles
(e.g. islands.esm.js); no vue/dist directory is left anywhere
else under blib.
5. git clean -xdf blib pm_to_blib Makefile Makefile.old && git checkout -- . to
clean up.
Signed-off-by: Paul Derscheid <paul.derscheid@lmscloud.de>
Signed-off-by: David Cook <dcook@prosentient.com.au>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
The JS variable base_url on barcode-print is never referenced. The remaining JS block did not contain TT interpolation, so data-tt-allowed can be removed.
Signed-off-by: Tomás Cohen Arazi <tomascohen@theke.io>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
In order to tidy all the script tags of the .tt files we isolate the
interpolation of Template::Toolkit variables in a separate script tag.
This patch only adds 'data-tt-allowed="true"' when needed. The work to
move the JS variable to a separate script tag has been done already on
bug 41324.
Signed-off-by: Jan Kissig <bibliothek@th-wildau.de>
Signed-off-by: Tomás Cohen Arazi <tomascohen@theke.io>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
While 'Options' might give the impression that the selection is
limited to the listed values, 'Examples' seems the better wording.
Signed-off-by: Katrin Fischer <katrin.fischer.83@web.de>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
To be more flexible we could allow the top-level domain instead of
having a predefined list.
However this patch still list the different available options (and adds
Brazilian) for easy selection.
Note that this system preference is not used to retrieve the image but
only to link to from the image. Contrary to what the reporter of the bug
report thought. And actually what I thought as well until I finished to
write the patch and started testing it!
Anyway now the patch is written, there should not be any changes in
behaviour. Most of the time the links are actually broken when not using
.com. So I think we either broke the whole feature at some point, or it
never really worked. On a separate bug we could eventually have a look
at fetching from the locale tld.
Test plan:
Turn on OPACAmazonCoverImages and AmazonCoverImages
http://localhost:8081/cgi-bin/koha/catalogue/detail.pl?biblionumber=5http://localhost:8080/cgi-bin/koha/opac-detail.pl?biblionumber=5
=> You see the PBP cover
click on it, it opens the lightbox, click "(see the original image)" at
the bottom
=> This is where the syspref AmazonLocale/AmazonLocaleTld acts
Modify the syspref's value and confirm that the link is modified and
point to the correct domain (leading to 404 or broken pages most of the time,
at least in my tests).
Signed-off-by: David Nind <david@davidnind.com>
Signed-off-by: Jonathan Druart <jonathan.druart@bugs.koha-community.org>
Signed-off-by: Katrin Fischer <katrin.fischer.83@web.de>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Test plan:
Run the updatedatabase script and notice that the sidebar is visible in
the circulation module but that you can hide/collapse it on demand.
Signed-off-by: David Nind <david@davidnind.com>
Signed-off-by: Emily Lamancusa <emily.lamancusa@montgomerycountymd.gov>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Where a patron had multiple overdues that ought to cause debarment, it was found that removing one removed the restriction.
This is because has_restricting_overdues() was looking up the ItemType object tied to the overdue issue, then passing it down to Koha::CirculationRules->get_effective_rules(). get_effective_rules(), however, expects an itemtype code.
This resulted in false negatives: the patron was treated as having no restricting overdues left where they in fact did.
Amend this by calling effective_itemtype() instead.
Note on side effects: I think this also will activate the previously potentially dead dedup logic in has_restricting_overdues (eq had been used to compare obj instance to obj instance which had no effect (itemtype() runs a find() on each iteration), and is now used to compare itemtype code to itemtype code).
Test plan (based on Pedro's comment number 949):
0) Pick a test opac patron, if using Henry from above remove the permissions
1) http://localhost:8081/cgi-bin/koha/admin/categories.pl, edit the test patron's category, confirm 'Overdue notice required' is set to 'Yes'.
2) Set syspref AutoRemoveOverduesRestrictions to when_no_overdue_causing_debarment.
3) http://localhost:8081/cgi-bin/koha/admin/circulation_triggers.pl, 'Add new trigger'. Leave 'Library' as 'All libraries' and 'Patron category' as 'All categories', set 'Item type' to 'Books'. For 'Letter' pick ODUE, check 'Email' under 'Transport type(s)', set delay to 5 days, for 'Restricts checkouts' select 'Yes', 'Submit'.
4) Check out barcodes 39999000019186 and 39999000007558 to the test patron (confirm both are itemtype 'Books'). Backdate both due dates 6+ days overdue.
5) perl misc/cronjobs/overdue_notices.pl -n — restricts the patron.
6) Check the patron's 'Restrictions' tab: confirm restricted.
7) Return barcode 39999000019186 only, leave 39999000007558
8) Check the patron's 'Restrictions' tab again. Notice they are still restricted (39999000007558 is still overdue).
9) Return barcode 39999000007558 also.
10) Check the patron's 'Restrictions' tab again. Notice the restriction is now removed (a hard refresh may be needed) as no restricting overdues are left.
Assisted-by: Claude Opus 5 (Anthropic)
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Since the decision was made that the add / edit trigger form should
never let a trigger be created or edited to not have a delay, it
follows that this must apply to the rule set reset flow as well.
Retain the warning message to the user, and also disable the
'Confirm reset' button if the reset would result in delay-less
sets.
The same applies to the trigger deletion functionality: if deleting
a trigger that is a default for all libraries would then leave any
libraries with one or more delay-less rule set, block the trigger
deletion (deactivate 'Confirm deletion') and inform the user. Here
as well we also still display the warning message.
Test plan:
A) Create one or more 'Default for all libraries' triggers.
B) Create one or more library-specific triggers.
C) For one or more of these, add itemtype and/or category specific overrides with not delay override.
D) For one or more of these, add itemtype and/or category specific overrides with a delay override.
E) Select 'Reset' on a rule set with overrides of type D) and notice that the 'Confirm reset' button is enabled.
F) Select 'Reset' on a rule set with overrides of type C) and notice that the 'Confirm reset' button is disabled,
and that the warning message explains why reset is unavailable
G) Select 'Delete' on the highest 'Default for all libraries' trigger, and notice that, where one or more library-specific overrides are of type D), 'Confirm deletion' is disabled, whereas it is otherwise enabled.
H) Delete a trigger for which no override exists and confirm success.
I) Delete a trigger for which only type D) overrides exist and confirm success.
J) Reset a rule set for which only type D) overrides exist and confirm success.
K) Reset a rule set for which no override exists and confirm success.
Assisted-by: Claude Opus 5 (Anthropic)
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
If the 'Delay' input field value is deleted using backspace or other,
block form submission. The same behaviour apply if the delete button
is used, as both clear out the value.
As triggers with no delay are not processed and would also cause
higher triggers to then also not be processed, we do not allow the
creation of triggers without a set delay.
Triggers when edited individually do not check for impact on any overrides
they might have. It is therefore safest not to allow submission of a trigger
where the user explicitly unset the value.
An explicit 'reset to fallback' feature is possible and would constitute a follow up enhancement.
Test plan (based on comment 872):
A) 'Administration' > 'Circulation triggers' > 'Add new trigger' for any library/category. Set a Delay, a Notice, and a transport type.
B) Select the Delay field's value and clear it via the 'X' clear button.
C) Notice 'Submit' is disabled.
D) Select the Delay field's value and clear it via the keyboard (Backspace/Delete). Click 'Submit'.
E) Notice 'Submit' is disabled.
Assisted-by: Claude Opus 5 (Anthropic)
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Looking at Bug 41167, this ought to fix the 404 error noted in comment 192 for package installs.
We do have a test asserting that RewriteRules must be carried over, so:
Test plan:
1. Apply patch
2. prove xt/check_apache_conf.t
3. Confirm it passes
That said, I don't think it is replicable in ktd - the copied file at /etc/koha/apache-shared-intranet.conf
will remain stale unless cp_debian_files.pl runs. The migration issue itself might be ktd-specific.
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Important note: the following assumes manage_circ_triggers is intended
as a sibling to manage_circ_rules, and that of the two, only
manage_circ_triggers ought to grant access to client-side circulation
trigger configuration actions.
When manage_circ_triggers was introduced, it was not applied client-side,
resulting in the router and menu permission checks incorrectly expecting
manage_circ_rules. Amend that.
Additionally, while the API endpoint blocked a user with only
manage_circ_triggers from editing non-triggers circulation rules,
it did not block a user with only manage_circ_rules from editing
trigger circulation rules. Add a guard clause.
Test plan (based on comment 872):
A) Go to a patron's 'More' > 'Set permissions' tab. Check 'catalogue' (needed for basic staff login), and under circulation-rules sub-permissions, check only 'Manage overdue notice triggers' - leave 'Manage circulation rules' unchecked.
B) Log in as that patron, go to 'Administration' - 'Circulation triggers' link present (admin-home.tt now correctly checks CAN_user_parameters_manage_circ_triggers), 'Circulation and fine rules' is not present.
C) Visit http://localhost:8081/cgi-bin/koha/admin/circulation_triggers directly - you are granted access.
D) Confirm that you can create and edit circulation triggers.
E) Go to a patron's 'More' > 'Set permissions' tab. Check 'catalogue' (needed for basic staff login), and under circulation-rules sub-permissions, check only 'Manage circulation rules' - leave 'Manage overdue notice triggers' unchecked.
F) Log in as that patron, go to 'Administration' - no 'Circulation triggers' link present (as we require manage_circ_triggers), however 'Circulation and fine rules' is present.
G) Visit http://localhost:8081/cgi-bin/koha/admin/circulation_triggers directly - you see "Error: You do not have permission to access this page" and are prompted to log in as a different user (as we require manage_circ_triggers).
H) Go to http://localhost:8081/cgi-bin/koha/admin/smart-rules.pl, and confirm that you can edit circulation rules for your library (no regression).
Assisted-by: Claude Opus 5 (Anthropic)
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
The following issues were identified and resolved:
- items.itemtype does not exist -> swap to items.itype
- leftover WHERE clause still referred to overduerules -> removed
- with the join on overduerules removed there is no need to GROUP BY anymore
- Koha::CirculationRules->all -> use next() instead
Test plan (as per comment 872):
A) In KTD, set 'TalkingTechItivaPhoneNotification' to 'On'.
B) Run perl misc/cronjobs/thirdparty/TalkingTech_itiva_outbound.pl --type=OVERDUE -o /tmp/out.txt -v
C) No errors / the script does not die mid run.
Quick disclaimer: I am not overly familiar with that script:
while I did remove the causes behind the script dying mid run,
I won't have spotted issues with expected output / outcome.
Assisted-by: Claude Opus 5 (Anthropic)
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Other complaints are invalid:
ContextA spelling - this is a variable
Use of _("") - this is in regex for 'overdue_(' when checking rules
git manipulation - not ideal adding and removing in same set, but this is a long
work, not asking for a squash here
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
- Display new manage_circ_triggers permission (add to template)
- Ensure loading messages are informative to the user
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
New tests:
- regression: mtt cells render as .fallback (bold italics)
- bg-success-subtle highlight on completed steps in the add flow
- disabled-Delete states:
- patron category filter set
- item type filter set
- default library, if any library has a trigger higher that default's highest
- redirect to admin home if not permitted to access the page
- in the edit/add stepper form
- check row actions are enabled at the 'selectOrAdd' step
- check row actions are disabled once add mode is entered
- check the action column is always available on the main page
Clean up / general updates:
- replace the /api/v1/** catch-all intercept with per-endpoint intercepts so @get-rules waits resolve reliably
- remove now-redundant cy.wait("@get-rules") calls
- drop unused mount / dayjs / dates imports
- ensure expected strings match UI where strings have changed
Assisted-by: Claude Opus 4.7 (Anthropic)
AI used to draft enhancements tests. Changes were reviewed, tweaked, and commited manually and gradually (with a pre-submission commit squash). Tests were run multiple times and adjusted iteratively to match code changes, cover regressions as those were found, and handle vue page load.
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Trigger deletion and rule set reset
- warn the user that resetting or deleting a given set will break sets using it as fallback.
- indicate which sets will be affected.
- display a preview of the post reset state of each affected rule set.
- preparation: allow TriggerTable to take a title parameter.
- refactor to remove 'has_rules' from the store's rule suffixes list so those match the DB.
- DRY: refactor to implement TriggerTable rather than use custom templates.
- select context stage handles narrower viewports
- prevent crash on missing rule set
TriggersTable: refactor: the <a> elements used as buttons are now button elements
- this is so they may be marked as disabled.
Main page - trigger actions (delete button display)
- disable the trigger deletion button.
- if the context has been narrowed down ('delete' deletes all explicit rule set per trigger per library).
- if the trigger belongs to default library AND one or more libraries have a trigger of a number bigger than the selected one.
This is to protect trigger sequentiality from being broken for the relevant libraries.
We also include an on hover message to explain the disabled button.
Add/ Edit trigger form
- clarify letters 'undo' button behaviour:
- swap the native vue select X for a fa-undo icon.
- display 'undo override and reset to fallback' on hover.
- only display the add/edit button at the selectOrAdd step.
- restore selected trigger yellow highlight (regression fix).
- display any temporarily saved data from prior step in a green 'success' container.
- only messages and data pulled from the DB for informative purposes display in blue containers.
Other
- redirect to correct URL (admin home) if store load fails
- mtt fallbacks display as such consistently (bold italics) - is a bugfix.
- catch errors on store initialisation
- remove whitespace duplicate
- ensure that mtt, if set, always display
Assisted-by: Claude Opus 4.7 (Anthropic)
AI used to help trace back potential bug sources, flag potential regressions, and draft enhancements. Changes were reviewed, tweaked, and commited manually and gradually (with a pre-submission commit squash).
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
We have nothing to do with it but since the patches change the file, it
trips the QA script.
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Fix: handle exiting modal
- prevent data reload when the user exits the modal
- as preparation for the above: refactor to use one handler only
Fix: new trigger saves with delay defined
- Identified an old refactor omission - we were still passing 'i' despite the from now relying on this.triggerNumber
Fix: return to correct trigger tab after save or reset
Fix: restore delete button and bold italic info
- restore display, fixes a regression introduced by the previous UI/UX patch set
Fix: order dropdown options by description or name (not code)
Fix: circ rule resets
- restore reset funtionality
(mtt was still occasionally passed as a string, and would then get compared to an array. Ensure this is not the case.)
- do not trigger the data refresh if the current ruleset did not get deleted
- make sure triggerNumber is defined when navigation is caused by reset
- add the notice code in brackets to letter names for display (eg. Overdue notice (ODUE))
Assisted-by: Claude Opus 4.6 (Anthropic)
AI used to help trace back potential bug sources and draft fixes. Changes were reviewed, tweaked, and commited manually and gradually (with a pre-submission commit squash)
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
- Refactor (prepation): Extract form sub-components for better maintainability
- Template stucture improvements (eg. consistent fieldset use)
- Ensure triggerNumber is always cast to integer from query params
- Add/Edit trigger form UI/UX improvements/fixes:
- Restore add/edit button functionality on trigger tables in the form
- Update Context section title to dynamically change depending on whether editing or displaying
- Display trigger context as informational when not in confirmContext mode
- Hide required labels and validation on view-only context fields
- Always display trigger context at the top of the form
- Add renewal/autorenewal summary (renewals allowed, automatic renewal) to circulation rules summary panel
- Sort notice letter dropdowns by name rather than code
- Update Circulation rules summary title to be more descriptive
- Scroll behaviour improvements:
- Extract scrollToElementById to store with polling and null guard
- Block page scroll when modal is open
- Disambiguate duplicate trigger-table element ids (trigger-table-main
vs trigger-table-form) to fix unpredictable scroll-to behaviour
Assisted-by: Claude Sonnet 4.6 (Anthropic)
Experimenting a little here - not quite certain where we draw the line for substantial use so I thought I would add a description in case it's useful, especially since the AI code guidelines are quite fresh:
AI was used primarily to help parse longer Vue templates (CirculationTriggersFormAdd.vue) and help isolate elements for extractions as well as highlight template formatting inconsistencies and propose improvements. The refactoring itself was achieved starting from AI-proposed changes which were then thoroughly re-worked manually towards the code changes submitted.
I also trialled using it (currently experimenting with this) for minor coding tasks where there is a clear precedent (for example, to add renewal/autorenewal information and do so in a way that follows exactly what we had written).
In those cases, each code change made by AI was the outcome of multiple iterations with direct feedback (which it did require a fair amount of) from me to ensure the output matched what I would have written myself.
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
The information message showing in the reset modal
still referred to an earlier version of the UI.
Update it to correctly name the relevant filter.
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
This is needed to handle creating new triggers
for libraries for which all earlier triggers are
defaults (do not exist in the DB , are not explicitly set)
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Replaces the broad manage_circ_rules permission requirement on the
circulation triggers page with a new dedicated manage_circ_triggers
sub-permission (under parameters, flag 3).
- Add manage_circ_triggers to userpermissions.sql
- Update atomicupdate to add the permission, grant it to existing
edit_notice_status_triggers users, and remove the old permission
- Update admin/circulation_triggers.pl to require manage_circ_triggers
- Add manage_circ_triggers as an OR option on all circulation_rules
API endpoints (GET/PUT /circulation_rules, /kinds, /config)
- Restrict manage_circ_triggers users in the API controller to
overdue_* rule names only on write
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
The Transport Type options when displaying in 'edit' mode were showing as
a comma separated string rather than as separate options (as they do when being set).
Ammends this so that they display as separate options consistently.
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
When editing a specific rule (e.g. */Kids/Books) alongside an all/all/all
rule, the conflict check incorrectly reported a conflict. The effective=true
API response always returns context as */*/* (fallback), so ruleSetInDb.context
differed from currentRuleSet.context (which was overridden to the real context).
Fix by overriding ruleSetInDb.context before the hasConflict comparison,
consistent with the existing pattern in setCurrentRuleSet.
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
The x-koha-authorization was using 'parameters: catalogue' which the
Koha auth system interprets as a 'catalogue' subpermission within the
'parameters' module (which doesn't exist), causing a 403 for any user.
The correct format for the top-level 'catalogue' permission is
'catalogue: "1"', consistent with all other read-only API endpoints.
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
- Add OR-type x-koha-authorization to /circulation_rules, /circulation_rules/kinds,
and /circulation_rules/config endpoints so users with either
circulate:circulate_remaining_permissions OR parameters:manage_circ_rules
can access them
- Enforce manage_circ_rules_from_any_libraries in the set_rules controller:
users without this permission may only set rules for their own library and
cannot set default (*) rules
- Add explicit use C4::Auth qw(haspermission) import to CirculationRules.pm
- Lock the library selector in the Vue UI for users who lack
manage_circ_rules_from_any_libraries, defaulting them to their own library
- Update set_rules() tests to cover the new permission restrictions
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
1. Letter context: setFilteredLetters() was reading library_id from
currentRuleSet (which has no top-level library_id property) instead
of this.context.library_id, causing library-specific letter templates
to never match.
2. False conflict on edit: conflict check fetched rule set from DB with
effective=false (raw), but the in-memory snapshot was fetched with
effective=true (with fallbacks). Comparing raw vs effective always
triggered a conflict when a context inherited any value from a more
general rule. Fixed by using effective=true for the DB fetch too.
3. Incomplete deletion: updateCircRuleSets() inside deleteRuleSet() was
called without await, creating a race condition where the page refresh
could complete before the deletion API call, leaving stale data visible.
Also added a guard to skip rule sets with no explicit rules for the
trigger being deleted, and simplified rulesForDeletion construction.
4. No rules displayed on initial load: the route watcher in
CirculationTriggersList fired immediately before init() had loaded
patron categories and item types, causing setAllExhaustiveEffectiveRuleSets()
to produce an empty result. Fixed with a metaInitialized flag that the
watcher waits for before proceeding.
5. DefaultToLoggedInLibraryOverdueTriggers syspref not respected: the
default_view template variable was computed but never applied to the
store's currentLibraryId. Fixed by passing it to init().
6. UI colors: info box background changed from #84c1f3 to #DAECFB (lighter,
matching the circulation rules page). Error alert messages changed from
alert-info (blue) to alert-warning (yellow) to match Koha conventions.
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
The config endpoint was gated on parameters: manage_circ_rules, while the
actual GET/PUT /circulation_rules endpoints require circulate:
circulate_remaining_permissions. Any user who can read/write circulation
rules should also be able to load the config that drives the UI.
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Sponsored-by: Glasgow Colleges Library Group <https://library.cityofglasgowcollege.ac.uk>
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Permissions.t:
- Remove edit_notice_status_triggers from userpermissions.sql so fresh installs
no longer seed this deprecated permission (existing installs are handled by
the atomic update)
- Add a defensive DELETE in the test to cover DBs that pre-date the atomic update
CirculationTriggers_spec.ts:
- Fix getMockCirculationRule: the helper was returning the old {rule_name,
rule_value, branchcode, ...} format; update to the current API format
{context: {...}, overdue_N_delay: ..., ...}
- Remove assertions for UI elements that do not exist (success/error flash
messages, "Existing triggers" heading, "rule set will be removed" text)
- Fix reset confirmation text to match actual component output
- Fix duplicate patron_category_id "ST" in mock patron categories
- Remove aspirational tests that checked non-implemented behaviour
- Add describe block for the no-default-triggers warning (no rules anywhere,
rules only at specific libraries, library link switches the filter)
- Add regression test for NaN trigger number when adding first rule for a library
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
When viewing the default library context with no triggers defined, show
a warning message explaining defaults apply to all libraries. If any
libraries have specific triggers, list them as clickable links that
switch the filter to that library. If no rules exist anywhere, show a
prompt to add the first trigger.
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
When no circulation rules exist for a library, triggerCounts[library_id]
is undefined, causing undefined + 1 = NaN. Use nullish coalescing to
default to 0 so the first trigger correctly gets number 1.
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
has_rules was a purely frontend marker stored as a circulation rule to
signal that a context had explicitly-saved overdue trigger rules. The
backend never reads it; the information is fully derivable from whether
any of delay/notice/mtt/restrict are non-null for a given context.
Replace all has_rules DB reads/writes with a hasExplicitRulesForTrigger()
helper that checks the four real rule suffixes directly. has_rules remains
as a computed in-memory concept in the Vue store (derived in findEffectiveRule)
so templates require no changes.
Also adds an atomicupdate to delete existing has_rules rows from the DB,
and removes the has_rules insert from the bug_10190 migration script.
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
- overdue_notices.pl: fix verbose sprintf - remove spurious \$triggered
argument (CLI 0/1 flag had replaced delay value in output) and use
the \$borr string that was computed but never included in the message
- overdue_notices.pl: remove %already_queued hash which was populated
in _enact_trigger but never read; dead code since the refactor
- C4/Overdues.pm GetBranchcodesWithOverdueRules: guard the global-rule
expansion with @branchcodes check so an empty result (no rules at all)
correctly returns an empty list instead of all branches
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
The config endpoint is used exclusively by the Circulation Triggers
Vue app, which is gated on parameters/manage_circ_rules. Using
acquisition/'*' and erm/1 was inconsistent copy-paste from an
unrelated endpoint and would 403 circ-rules admins while granting
access to unrelated module users.
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
- Patron.pm has_restricting_overdues: sort ASC not DESC so the oldest
(most overdue) item is checked per branch/itemtype; DESC caused the
deduplication skip to discard the very items most likely to trigger
a restriction
- overdue_notices.pl: fix --itemtypesout option name in GetOptions
(was 'itemtypeouts', never matched the documented --itemtypesout flag)
- overdue_notices.pl: fix typo 'catagorycode' -> 'categorycode' in
--borcatout search, which silently broke category exclusion
- circulation_triggers.pl: add flagsrequired => parameters/manage_circ_rules
so the admin page enforces the correct permission; also add missing
use Koha::Notice::Templates
Signed-off-by: Michaela <michaela.sieber@kit.edu>
Signed-off-by: Nick Clemens <nick@bywatersolutions.com>
Signed-off-by: Pedro Amorim <pedro.amorim@openfifth.co.uk>
David Cook
Paul Derscheid
Lucas Gass
Jan Kissig
Jonathan Druart
Pedro Amorim
Katrin Fischer
Chloé Zermatten
Nick Clemens
Victor Grousset/tuxayo
Martin Renvoize