`report_journal_ledger` is auto-vacuumed as any transient model, but has some
ondelete="cascade" constraints that auto-remove subtables when a record is removed,
doing this operation very slow when selecting these sub-records.
Letting default ondelete="set null" would result in same performance bottleneck,
as the select on sub-table is performed the same for setting "null" value on them.
As a solution, and for avoiding a costly index operation, we delete by SQL sub-tables
rows in advance.
A bit of extra logic has been added for avoiding to remove that records if it's not
the turn of vacuum the parent table.
* Passing values to general ledger was stripping some correct records
* Computed field for hiding lines doesn't have proper dependencies nor is not
taking into account float currency accuracy
All reports:
- Rename field to hide accounts at 0 to 'hide_account_at_0'
Trial Balance:
- Add possibility to filter by hierarchy levels
- XLSX format will show the hierarchy levels in bold
General Ledger:
- Add the possibility to filter by analytic tags
- Fixes an error on the default date
Journal Ledger:
- The filter on Journals is now optional. If the user does not choose
a journal, by default it will display all journals.
Aged Partner Balance:
- Fixes an error on the default date
of undistributed profits/losses account, and provide a footer note
to the user explaining why will the ending balances will not be zero, but the
period's total profit and loss.
- Adds 'Period balance' column
- Renames the option 'Hide accounts at 0'. Means no initial, no debit, credit
or ending balance for the period.
- Fixes logic to remove lines with 0 activity for the period.
When there are a lot of account.move.line (several millions) and print any of
the Qweb reports, that will generate also a lot of transient objects.
As these objects are created with an "insert" query, the cleaning normally
triggered by the count of the records in transient tables is not done, so only
the cleaning based on the age of the records is processed (by default, records
older than 1 hours are deleted), but the cron task is only ran one time per
day. For large setups this can lead to memory errors at that point. This change
prevents the memory error by executing the transient record cleanup for the
report models in this module in SQL.
When this module is installed along with other chart account different from generic one,
the number of expected accounts and the computation change (for example, in Spain,
the unaffected earnings account is 129000, choking with group with code prefix 1).
This commit makes the tests resistent to these changes.
Update indentation, remove empty lines from header.
Update test.
Update pylint.
Remove company_id on computing accounts, since account.group is not a company based model, filtering accounts is done on trial balance report.
Update account variables.
Improve condition in padding on accounts.
Add option to print hierarchy based on defined accounts/computed accounts.
Add VAT report, hierarchy from tax tags ans taxes.
Fix pylint, xlsx report generation header.
Update code to select code_prefix or name.
Update code to select code_prefix or name.
Update code to select code_prefix or name.
Fix domain in base amounts in vat report.
Change trial balance code_prefix or name.
Update trail balance, add tests for vat report.
Update pylint, amounts as monetary, many2one option on generation excels.
Update pulint.
Add VAT Report in readme.
Add VAT Report in readme.
Update array_agg.
Update array_agg.
Update array_agg.
Add option in VAT Report to be printed on Tax Tags - Tax Groups.
Add widget to hierarchy_on on trial balance.
sbejaoui
Pedro M. Baeza
Iryna Vushnevska
Lara Baggio
Jordi Ballester Alomar
Pedro M. Baeza
mreficent
Abraham Anes
Andrea
Patrick Tombez
mpanarin
hveficent
Fekete Mihai
rgarnau
Adrien Peiffer (ACSONE)