20151204 (Friday, 04 December 2015)

Selecting photos from Shotwell using their date

Annely asked me to send all my 2015 photos about our folk dance group to Hillar. Since this is my second year with them and I don’t want to manually fiddle with photo files, I added two options –after and –before to shotwell2blog

Learning by doing: logrotate

I saw that #647 is still not solved. Let’s see what’s wrong:

$ sudo logrotate -v /etc/logrotate.d/lino
reading config file /etc/logrotate.d/lino

Handling 1 logs

rotating pattern: /var/log/lino/prod/system.log  weekly (52 rotations)
empty log files are not rotated, old logs are removed
considering log /var/log/lino/prod/system.log

error: skipping "/var/log/lino/prod/system.log" because parent
directory has insecure permissions (It's world writable or
writable by group which is not "root") Set "su" directive in
config file to tell logrotate which user/group should be used for

So I added this line to the config file:

su root www-data

Yes, now it looks better:

$ sudo logrotate -v /etc/logrotate.d/lino
reading config file /etc/logrotate.d/lino

Handling 1 logs

rotating pattern: /var/log/lino/prod/system.log  weekly (52 rotations)
empty log files are not rotated, old logs are removed
switching euid to 0 and egid to 33
considering log /var/log/lino/prod/system.log
  log does not need rotating
switching euid to 0 and egid to 0

And finally I ask to force the first rotation (otherwise I’d have to wait again one week before seeing whether it really works):

$ sudo logrotate -f /etc/logrotate.d/lino

Basically keep in mind: logrotate FILENAME simply executes the rotation(s) specified by the given config file FILENAME. The daily cron job does exactly this every morning for every config file. The -v option adds verbosity. The -f option means “rotate the file even if you think it is not necessary”.

Wildcard in column_names

There is still one subtle problem, I think it was already before yesterday (thus not caused by #650):

File "docs/specs/users.rst", line 79, in users.rst
Failed example:
    ==== ================= =================
     ID   Author            User
    ---- ----------------- -----------------
     1    Hubert Huppertz   Theresia Thelen
     2    Alicia Allmanns   Theresia Thelen
     3    Mélanie Mélard    Theresia Thelen
    ==== ================= =================
    ==== ================= ================= ========== ====================
     ID   Author            User              Workflow   Description
    ---- ----------------- ----------------- ---------- --------------------
     1    Hubert Huppertz   Theresia Thelen              *Authority object*
     2    Alicia Allmanns   Theresia Thelen              *Authority object*
     3    Mélanie Mélard    Theresia Thelen              *Authority object*
    ==== ================= ================= ========== ====================

The column_names of users.Authorities has a wildcard.

The two standard virtual fields “Workflow” and “Description” should be visible only when explicitly specified. But it should be possible to add them manually.

Welfare test suite

The Welfare test suite still has a few failures after #650.

Endbenutzer können keine SEPA-Konten sehen

Gerd reported that they cannot see any imported account. This was due to a bug introduced by renaming field last_movement to last_transaction:

  File "/python2.7/site-packages/django/core/handlers/base.py", line 132, in get_response
    response = wrapped_callback(request, *callback_args, **callback_kwargs)
  File "/python2.7/site-packages/django/views/generic/base.py", line 71, in view
    return self.dispatch(request, *args, **kwargs)
  File "/python2.7/site-packages/django/views/generic/base.py", line 89, in dispatch
    return handler(request, *args, **kwargs)
  File "/lino/lino/modlib/extjs/views.py", line 543, in get
    for row in ar.sliced_data_iterator]
  File "/lino/lino/core/store.py", line 1094, in row2list
    v = fld.full_value_from_object(row, request)
  File "/lino/lino/core/store.py", line 364, in full_value_from_object
    return self.vf.value_from_object(obj, ar)
  File "/lino/lino/core/fields.py", line 476, in value_from_object
    return m(obj, ar)
  File "/lino-cosi/lino_cosi/lib/sepa/models.py", line 121, in statements
    return ar.obj2html(b2c, dd.fds(b2c.last_movement))

201512-04 11:02:47 WARNING ajax : AjaxExceptionResponse AttributeError: 'Account' object has no attribute 'last_movement'

lino_xl.lib.b2c.fixtures.demo now connects the important accounts to some random partner.

New tested document Import bank statements (SEPA BankToCustomer).

Rumma & Ko is going into the right direction

Nice: Susan Ward, Canadian Small Business Expert, writes in The Advantages of Being a Contractor that Self-Employment is the Future and that “according to a study by Intuit more than 80 percent of large corporations plan to substantially increase their use of a “flexible” workforce. The study predicts that by 2020 “contingent” workers will exceed 40 percent of the U.S. workforce.”

Which seems to indicate that my visions for Rumma & Ko with Lino are aiming the right direction.

Released Atelier 0.0.16

This was necessary because Lino was failing with Atelier 0.0.15 which still tried to import the importlib coming with Django. See Changes in atelier.

Cannot delete a Place although it is not used

There is another occurence of #418 in Eupen. It says:

IntegrityError (1451, ‘Cannot delete or update a parent row: a foreign key constraint fails (dbname.`contacts_person`, CONSTRAINT city_id_refs_id_d349fc28 FOREIGN KEY (city_id) REFERENCES countries_place (id))’)

Last time (20150824 (Monday, 24 August 2015)) I

The system.log file shows that Mélanie was tidying up to remove duplicates. Before deleting place 1062 she had changed a few address rows to point to the other instance. After the error message she changed the name of place 1062 to “obsolet”. Which is okay for them. But why do Lino and MySQL disagree about whether there are still other rows referring to place 1062?

I tried to create a new place (1063) and then to merge the place 1062 into it. Also here Lino does not see “anything related” to that place:

Sind Sie sicher, dass Sie obsolet nach test fusionieren wollen?

obsolet wird gelöscht

Here is again their MySQL engine status:

mysql> show engine 'innodb' status;
151204  9:56:43 Transaction:
TRANSACTION 116B0C9, ACTIVE 0 sec updating or deleting
mysql tables in use 1, locked 1
15 lock struct(s), heap size 3112, 8 row lock(s), undo log entries 1
MySQL thread id 1062531, OS thread handle 0x7f34f0212700, query id 111621549 localhost django updating
DELETE FROM `countries_place` WHERE `id` IN (1062)
Foreign key constraint fails for table `dbname`.`contacts_person`:
  CONSTRAINT `city_id_refs_id_d349fc28` FOREIGN KEY (`city_id`) REFERENCES `countries_place` (`id`)
Trying to delete or update in parent table, in index `PRIMARY` tuple:
DATA TUPLE: 11 fields;
 0: len 4; hex 80000426; asc    &;;
 1: len 6; hex 00000116b0c9; asc       ;;
 2: len 7; hex 52000052691801; asc R  Ri  ;;
 3: len 6; hex 20457570656e; asc  Eupen;;
 4: len 2; hex 4245; asc BE;;
 5: len 0; hex ; asc ;;
 6: len 0; hex ; asc ;;
 8: len 0; hex ; asc ;;
 9: len 0; hex ; asc ;;
 10: len 0; hex ; asc ;;

But in child table `dbname`.`contacts_person`, in index `contacts_person_b376980e`, there is a record:
PHYSICAL RECORD: n_fields 2; compact format; info bits 0
 0: len 4; hex 80000426; asc    &;;
 1: len 4; hex 800311ce; asc     ;;

I tried the following in a Django shell:

from lino.api.shell import *
for p in contacts.Person.objects.filter(city_id=1062):
    print p
print "Done"

The output is simply “Done”. No persons refer to that place.

Okay there other types of database objects that might refer to a city (see:ref:welfare.specs.ddh), but obviously Lino would not see them, otherwise it would have told so before asking to delete the record).


Renamed “Lino Faggio” to “Lino Voga”

When choosing “Faggio” as name I did not know that there are real persons living in Italy whose name is Lino Faggio. I prefer to change the name before the product gets more widely known…

Upgrade with surprises

The online demos were broken for about 1 hour because I had some surprises after upgrading. #60 had caused a number of bugs that had sneaked across our test suite.

And BTW the polls demo is still broken.