Tuesday, August 6, 2019

Accounting

I continued entering our accounting into Lino Così. I am still fiddling on EKR 7/2018 which is our first case of internal clearing (#3106).

How it looked in TIM:

Accnt  Z Bezeichnung                    Debit      Credit     *
──────┬─┬──────────────────────────────┬──────────┬──────────┬─
440000│E│Eesti Energia AS              │          │     79.97│*
440000│E│Eesti Energia AS              │     79.97│          │*
440000│E│Saffre Luc                    │          │     79.97│*
614100│ │Elekter                       │     23.33│          │
472200│ │Deklareerimata käibemaks saada│      4.66│          │
440020│ │Hankijad (arveldatud töötajate│     51.98│          │*
      │ │Saldo D=                      │    159.94│    159.94│

I decided to introduce a few subtle changes compared to TIM. While TIM booked the third movement using the suppliers general account, Lino books it into a new account “Internal clearings”. And I map account 440020 to 600020 (i.e. we classify it as a costs account rather than a credits account). How it looks in Lino:

Account

Partner

Debit

Credit

Match

Cleared

440000 Suppliers

Eesti Energia AS

79.97

EKR 7/2018

Yes

440000 Suppliers

Eesti Energia AS

79.97

EKR 7/2018

Yes

462100 Internal clearings

Saffre Luc

79.97

EKR 7/2018

No

472200 VAT deductible

13.32

Yes

600020 Costs paid for workers (to be paid back)

23.33

Yes

614100 Electricity

43.32 159.94

159.94

Yes

More optimizations en passant:

I fixed a bug in lino_xl.lib.ledger : when the Account.common_account was updated via the web interface, Lino did not “realize” this until the server process was restarted. Fixed. To do this without any hack, I added a new method lino.core.diff.ChangeWatcher.get_old_value().

inv prep test ended with error 245 in project book

This morning, before going to bed, I told my computer to pp -a welfare inv prep test. It worked its way through the demo projects, running manage.py prep --noinput --traceback in each of them. And then, out of nowhere, in lino_book.projects.lydia, (the 30th of the 46 demo projects), that command fails. It says:

CommandError: inv prep test ended with error 245 in project book

When I manually run the command which failed, it does not fail again.

What does exit code 245 mean? Does it have a special meaning? “When a command terminates on a fatal signal whose number is N, Bash uses the value 128+N as the exit status.” (gnu.org) See also tldp.org

245 - 128 = 117 and my question becomes “What means signal 117?”. I scanned the man page

Other people had this exit code. For examples here <https://stackoverflow.com/questions/10054073/failed-with-exit-code-245-xcode>`__ and here.`

But 256 - 11 = 245. So the number might just mean that signal -11 (Segmentation fault) was received. This can simply be caused by Low memory or Faulty Ram memory (link)

I cleaned up some places which caused warnings about unclosed resources.

getlino doesn’t start the database server

I noticed that getlino configure did not start the database server. Probably fixed. The DB_ENGINES list has a new new column service. There was maybe also a bug in getlino.utils.Installer.setup_database(), the runcmd was called with sudo -u postgres -c bash "{cmd}" but (imho) sudo -u postgres {cmd} is enough.

En passant I saw that on Ubuntu Bionic our getlino would fail:

Package libmysqlclient-dev is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source
However the following packages replace it:
  libmariadb-dev-compat libmariadb-dev

E: Package 'libmysqlclient-dev' has no installation candidate
E: Package 'mysql-server' has no installation candidate

Changing number of calendar entries in welfare demo project?

I “fixed” the following failure in Lino Welfare by adapting the expected output. Though I am quite sure that a few days ago I did exactly the opposite. I have no explanation yet.

File "docs/specs/weleup/eupen.rst", line 1033, in eupen.rst
Failed example:
    walk_menu_items('rolf', severe=False)
    #doctest: +ELLIPSIS +NORMALIZE_WHITESPACE +REPORT_UDIFF
Differences (unified diff with -expected +actual):
    @@ -12,5 +12,5 @@
     - Büro --> Meine Datenkontrollliste : 0
     - Kalender --> Meine Termine : 6
    -- Kalender --> Überfällige Termine : 81
    +- Kalender --> Überfällige Termine : 37
     - Kalender --> Meine unbestätigten Termine : 3
     - Kalender --> Meine Aufgaben : 1
    @@ -70,5 +70,5 @@
     - Konfigurierung --> Büro --> Ereignisarten : 11
     - Konfigurierung --> Büro --> Meine Einfügetexte : 1
    -- Konfigurierung --> Kalender --> Kalenderliste : ...
    +- Konfigurierung --> Kalender --> Kalenderliste : 12
     - Konfigurierung --> Kalender --> Räume : 1
     - Konfigurierung --> Kalender --> Regelmäßige Ereignisse : 16
    @@ -140,5 +140,5 @@
     - Explorer --> Büro --> Ereignisse/Notizen : 112
     - Explorer --> Büro --> Einfügetexte : 3
    -- Explorer --> Kalender --> Kalendereinträge : 358
    +- Explorer --> Kalender --> Kalendereinträge : 311
     - Explorer --> Kalender --> Aufgaben : 37
     - Explorer --> Kalender --> Anwesenheiten : 620
**********************************************************************
1 items had failures:
   1 of  26 in eupen.rst
***Test Failed*** 1 failures.

Accounting

lino_xl.lib.vat.InvoicesByJournal did not display the right sorting order. It was using [‘-id’] instead of the order defined in lino_xl.lib.ledger.ByJournal.

‘Duration’ object has no attribute ‘_text’

Tonis added a deconstructible decorator to the custom values in lino.utils.quantities, but this caused most tests to break with above error message.

After reading the Django docs I guess that @deconstructible doesn’t work as expected because these classes have a __new__() method (no __init__() method). I guess that we must simply write our own deconstruct method.

I reviewed the code and probably fixed the issue, added some cases to Quantities.

Running apt-get upgrade on travis

I added sudo apt-get update -y (and upgrade) to the .travis.yml files of book and welfare because a failure suggested it.