Re: [Gnumed-bugs] <bug>: Again, with the 2 files replaced...

[Karsten Hilbert]

Hello Marc,

thanks for persisting with us in tracking this down.

In the attached files I have

- adjusted the logging to produce less noise and hone in
  on the interesting list only

- made the code not fail early on use anymore because
  we now know the path of failure (not the exact reason
  just yet) such that we can maximize the logged
  information when the failing path is traversed until
  it eventually fails as it did initially ("no such
  item" on double-click)

- removed one entry into the provider inbox repopulation
  code (namely from when the notebook tab is raised) as that
  event does not need to reload the inbox (anymore) because
  any changes in the underlying data (patient change, docs
  change, test results change, explicit message received)
  will schedule a data-reload-on-paint *anyways* which will,
  of course, happen just when the notebook plugin is raised
  - no need for another *explicit* reload at that time

We'd added that explicit reload-inbox-on-plugin-raise when
we did not yet properly listen for database signals relevant
to provider inbox changes.

Please try to tickle the problem again. If you happen to
have a long run without error and you're getting the feeling
"I really should've seen the problem by now/this" then just
mail me a log of that lucky run even if the error did not so
far occur.

Thanks,
Karsten
-- 
GPG key ID E4071346 @ gpg-keyserver.de
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346

gmProviderInboxWidgets.py
Description: Text Data

gmListWidgets.py
Description: Text Data