20150915 (Tuesday, 15 September 2015)¶
I stumbled over #523 and investigated the reason. We must
probably just add param_values
to the inheritable attributes
to be copied to the other request (in setup_from
). I will not fix it now
since Hamza is working on #520. But I can write a test case
to reproduce it.
Links to tickets¶
I noticed that links to tickets didn’t work any more after 2015-09-11 because of #469. Fixed. That’s probably why Hamza didn’t find #521 yesterday. TODO: Note that I only fixed it quickly, I replaced:
extlinks.update(ticket=('http://bugs.lino-framework.org/ticket/%s', '#'))
by:
extlinks.update(ticket=(’http://bugs.lino-framework.org/tickets/Ticket/%s’, ‘#’))
in my blog’s conf.py
file. A better solution would be to add
a custom url pattern to the public UI of Lino Noi and to maintain
existing links functional.
Ticket #159¶
Steve managed to reproduce #159 (‘NoneType’ object is not iterable (after duplicating a budget)). It turns out that the symptom of this problem was the same as that of #471 (Become the author after duplicating a budget): after duplicating a budget, the author of the new budget should be the user who asked to duplicate it, not the original author.
I extended test_debts
to reproduce also this
problem.
A first subtle consequence for the framework: Changed the exception
handling during TableRequest.execute
. When an exception
occurs during get_data_iterator()
, then we want it to be
reported only once. This remains. But until now it was reported as a
direct call to logger.exception()
. Now it is being forwarded
(re-risen). Because the old behaviour was hiding the real location, a
reason why #159 had been so hard to find.
A second subtle consequence: the known_values attribute of a
TableRequest
had a
default value None, overriding the default value {}
already
defined in lino.core.requests.BaseRequest
. I guess that this
was not intended.
I also added some explanations in the docstring of
lino.core.model.Model.on_duplicate()
.
Ticket #523¶
Since anyway I did some changes to Lino today (and since I hope that
there won’t be any conflicts), I then fixed #523. Most time
went (as usual) into writing a test case (in
working : Work time tracking). The actual bug was in
lino.core.requests.BaseRequest.set_selected_pks()
. En passant
I also fixed the test suite (changes in tickets (Ticket management in Noi)).