• last updated 4 hours ago
Constraints
Constraints: committers
 
Constraints: files
Constraints: dates
Fix the case of the last comment

Improve performance of delete operation

Usability improvement: disable the buttons to flag/unflag artifacts when these are already flagged/unflagged (applies also to bulk buttons)

Behavior reform:

now commenting an artifact won't be considered as "OK". To flag an artifact as "OK" or "for review", one has to explicitly click on the green and red buttons.

Apply filters also to new images coming from the websocket

allow fractional minutes (useful for short quizzes)

Make the border indicating the status of an artifact's review slightly bigger

Implement buttons to set all artifacts for a user as ok or not ok

file review-all.tcl was initially added on branch oacs-5-10.

Improve robustness of upgrade script for legacy applications

In case, an installation has the legacy view (must be from very old installations)

in place, it is necessary to drop it before registered_users, otherwise

PostgreSQL will complain that it cannot drop view registered_users because

other objects depend on it.

Many thanks to Franz Penz to report this potential problem for old sites.

Make ad_html_security_check configurable

ad_html_security_check has now three optional attributes

to make it configurable for different situations

-allowed_tags

-allowed_attributes

-allowed_protocols

If these attributes are not specified, the behavior is exactly like

before. This change makes it also easier to regression test this

function, since the behavior does not necessarily depend on a site's

parameter settings.

The function was also modernized, new regression tests were added.

Use only ok and flagged as color schemes, so that results can be investigated also while review is in progress

Handle the case where only one type of artifacts is collected

Replace private with public api

Actually not enforce options validation on disabled formfields

Don't enforce options validation on disabled fields, demonstrate that this won't allow to insert invalid values in the field

Fix the test, which was not actually reproducing the issue

in order to do so, I had to fix the behavior of acs::test::xpath::get_form_values: we should in fact NOT extract the value of disabled formfield, as a real browser will not send them in the POST request.

Expose behavior:

extend date formfield test to include also a date in a specific format from downstream, set to disabled. Set the day of the month as a number < 10. Show that in this case the validation will complain about an invalid value. This because the field is represented as " 1" rather than the expected "1"

Reduce hard errors in the search indexer on invalid file content

This change uses util::file_content_check introduced with acs-tcl

5.10.1d9 to detect error situations before external programs are

called, which can lead to unpredictable error messages.

bumped version to 5.10.1d1

    • -3
    • +3
    /openacs-4/packages/search/search.info
    • -2
    • +2
    /openacs-4/packages/search/tcl/search-procs.tcl
fix issue shown with acunetix testing (without login)

    • -3
    • +3
    /openacs-4/packages/search/www/search.tcl
guarantee uniqueness of IDs when multiple aggregated forms are used in one page (e.g. exam protocol)

Constrain the language we return for the connection to the ones we support: could happen on system that have supposted a language in the past and still have users set it as their preferred language

Additional CSP exceptions for Xinha

improve class diagram drawing logic

- Draw always the superclass of the object of

interest, which might in the case of multiple

inheritance not in the "things" variable.

- don't draw lines twice in cases of multiple inheritance

include DB interface classes in API-doc

pass "hide methods" to svg generation

switch from xo::db::sql to acs::dc interface

Make the initial population of the request-monitor counters more robust

simplify logic

Fix variable name