Wednesday, October 21, 2015

I merged more changes by Hamza for #505 into master:

$ git fetch hamza
remote: Counting objects: 28, done.
remote: Compressing objects: 100% (28/28), done.
remote: Total 28 (delta 12), reused 0 (delta 0), pack-reused 0
Unpacking objects: 100% (28/28), done.
   5ba0ba8..85501a3  master     -> hamza/master

$ git merge hamza/master
Updating a038fbb..85501a3
 docs/tested/demo.rst           |  8 ++++----
 docs/tested/general.rst        |  5 +++--
 docs/tested/ledger.rst         | 58 +++++++++++++++++++++++++++++++++++++++++-----------------
 lino_cosi/lib/sepa/ |  8 ++------
 lino_cosi/lib/sepa/   |  6 ++++--
 lino_cosi/lib/sepa/       |  9 +++------
 6 files changed, 57 insertions(+), 37 deletions(-)

$ git push
Counting objects: 50, done.
Delta compression using up to 8 threads.
Compressing objects: 100% (28/28), done.
Writing objects: 100% (28/28), 3.26 KiB | 0 bytes/s, done.
Total 28 (delta 24), reused 0 (delta 0)
   a038fbb..85501a3  master -> master

For Welfare I still had to pull down Hamza’s changes after having confirmed and merged his pull request using the web interface, but Git surprised me:

$ git pull
remote: Counting objects: 16, done.
remote: Compressing objects: 100% (9/9), done.
remote: Total 16 (delta 13), reused 10 (delta 7), pack-reused 0
Unpacking objects: 100% (16/16), done.
   e6d739e..806f191  master     -> origin/master
There is no tracking information for the current branch.
Please specify which branch you want to merge with.
See git-pull(1) for details

    git pull <remote> <branch>

If you wish to set tracking information for this branch you can do so with:

    git branch --set-upstream-to=<remote>/<branch> master

As it seems (according to this), I somehow did not yet set my upstream master:

$ git branch --set-upstream-to=origin/master master

Atelier, Noi, Welfare and Cosi now watched on

I created three new projects at for Atelier, Lino Noi, Lino Così and Lino Welfare. Cheat sheet for creating a Drone project for a Lino application:

  • Environment Variables:

  • Commands:

    pip install -e .
    mkdir -p $LINO_CACHE_ROOT
    fab initdb test

I did this mainly because this seems the easiest way to automatically notify Hamza when I push a set of changes to my repos.

Maybe it won’t be trivial to get Cosi and Welfare watched on Drone because we test them with the latest development version of Lino (while on Drone they will use -of course- the latest released Lino version). It means that I will have to release Lino more often. But you are right, that’s actually not a big problem, and it is necessary to do it.

Lino Così then failed on

ImportError: Problem installing fixture '/home/ubuntu/virtualenv/python2.7/lib/python2.7/site-packages/lino/modlib/countries/fixtures/': No module named

… despite the fact that I specified in the tests_require of my file. I guess that the pip install -e implicitly causes pip to not install dependencies. But I don’t like to use a requirements.txt file because that’s redundant with the To be observed. #600.

Fixed some test failures in Welfare

I fixed some failures in Accounting for Lino Welfare which were caused by Hamza’s changes. Actually he wanted to look at them.

Since sepa now needs ledger, Hamza had to change the order of these plugins in get_installed_apps.

This caused lino_welfare.modlib.ledger.fixtures.demo to fail because it was now being loaded before lino_xl.lib.sepa.fixtures.demo. But the latter creates the contacts with an IBAN account number needed by the former. So I moved the creating of incoming demo invoices from lino_welfare.modlib.ledger.fixtures.demo to lino_welfare.modlib.sepa.fixtures.demo.

Another series of failures was due to the fact that demo user “wilfried” now gets created earlier (again due to the new sequencing order of the plugins) so that some user id’s have changed. Hamza, I leave this to you as an exercise!

When pushing my changes, I saw yet another problem: our demo file lino_welfare/projects/eupen/tests/COD_20150907_O25MMF107I.xml (which I added 20151013) had been deleted again. I first thought that this was also because Hamza was not notified about my commits and therefore did not merge them into his fork. But the explanation was easier: delete_imported_xml_files now works! And therefore the test_import_sepa test deletes our demo file each time when being run. Solution was to set the default value of this setting to False.

#595 (Tx25 : No handler for BurialModes)

I fixed #595. The problem itself was trivial (just add a handler in lino_welfare.modlib.cbss.tx25), but testing this type of issue is always a bit tricky because I have only confidential data for testing these.

This time it caused some subtle bugfixes with danger of side effects:

  • cbss.RetrieveTIGroupsRequest had a forgotten manual do_print action defined. Thiscaused Lino to continue printing Tx25 requests using DirectPrintAction instead of using the excerpts system.

  • And then the excerpt sdid not appear in the client’s history because there was no lino_welfare.modlib.cbss.mixins.CBSSRequest.get_excerpt_option().

  • The certificate.body.html template tried to extend base.body.html with a wrong path:

    {% extends "excerpts/Excerpt/base.body.html" %}