We have dropped support for the skiptest attribute for XML . It was
no longer used by any of the standard modules and we think it is always
better that the tests actually test the loading of all the data.
As it was pointed out in Invalid attribute "skiptest" (#8735) · Issues · Tryton / Tryton · GitLab , this will have a massive negative impact on GNU Health. Re-loading of large static lists is not really beneficial…
So probably better to load and update this data with a script, this will speedup the update process and will allow to update the data without requiring a gnuhealth update.
Also during development cycle, you can use the DB_CACHE environment variable to reuse an existing database backup. This will prevent to load the XML on each test run.
And finally it is the CI that runs the tests so if it takes a little bit more time, it is not really a problem compare to the benefit to ensure that all data can be really loaded.