- Jan 01, 2019
-
-
- Dec 31, 2018
-
-
- Dec 30, 2018
-
-
Currently, weboob commands still return 0 even when an exception is raised. Let bcall_errors_handler decide the return code instead of always 0. A good return code is hard because of several things: * there might be multiple exceptions in a BCallErrors, so values returned by bcall_error_handler (single exception) cannot easily be combined * the base implemententation of bcall_errors_handler cannot be polluted with a lot of exception subclasses So, any non-MoreResultsAvailable will cause a non-zero return code. Should fix https://git.weboob.org/weboob/devel/issues/74.
-
"boobank history" lacks an account argument but boobank returned 0 as if there was no error. Related to https://git.weboob.org/weboob/devel/issues/74. TODO: return error codes for BCallErrors
-
-
-
-
-
- Dec 29, 2018
-
-
Romain Bignon authored
-
This should minimize the human work required for a release
-
-
-
-
And avoid global variables
-
-
-
-
Using the current user and group does not make much sense. I would have put something other than root/root like weboob/weboob, but it only works if the user and group exists, despite them being stored as text in the tarball.
-
-
-
It was used by the old bash completion, but otherwise not very useful to end-users. Moreover, it was giving wrong results! For example, weboobcfg is not the actual callable application, it is weboob-config.
-
Since we already generate the man pages automatically, we can generate the bash completions too. It is much faster and does not risk making any network requests. fixes #194
-
-
An unfortunate commit. I am not aware of anyone still using a N900 or similar. The Hildon library is not maintained anymore. Blame Sephen Elop.
-
Since there is no way to provide build options, (the [] syntax seem to only work for dependencies) we use the environment variable WEBOOB_SETUP to allow people installing weboob from various tools to pass the options. This isn't a good solution, but it's better than nothing.
-
-
For some investments lists the ISIN codes are missing in the API, but we can fetch them on the invest detail page. When the concerned connection will switch to the new interface, I will check how SRD invests are organized and if we can fetch missing codes the same way. Closes: 8489@zendesk
-
Some connections were still bugging since the new website arrived, especially because of the "home page" with market data that can have several URLs. From the home page or the view page, I added an on_load() to go directly to the accounts page. Some connections are still using the former website and there are no accounts on the new website architecture so I re-implemented the former behavior. Closes: 8489@zendesk
-
Every time we had a couple having 2 market type account, we always retrieved the first one when iter_investing. Fixed by checking the full name of the owner, since you can only have one account of those type.
-
- Dec 27, 2018
-
-
Romain Bignon authored
-
- Dec 23, 2018
-
-
Romain Bignon authored
-
- Dec 22, 2018
-
-
linebourse API When there are several accounts on the market space we need to fetch to account_code for the account of interest before fetching its invests or history, otherwise we always land on the same page and we scrape the same invests for all accounts. Closes: 8683@zendesk
-
- Dec 21, 2018
-
-
simplejson has an internal way of formatting Decimal so it bypasses our formatter. Since weboob often deals with money, floats are not a good idea. Previousl: Decimal('1.1') => 1.1 with simplejson+WeboobEncoder Decimal('1.1') => '1.1' with json+WeboobEncoder Now both give '1.1'. Under simplejson: Using dump(cls=...) always initialize the Encoder class with the use_decimal=True argument (unless the argument is provided in the dump function, but that requires a lot of alteration in Weboob, and I would rather remove simplejson support altogether). So when present, we force it. Under json: There is no use_decimal, so we already go to the default() method of the Encoder.
-
-
users with only savings account can't go on 'CPTSYNT1' page When we go on 'CPTSYNT1' page, we are redirected on another page The page seems to not be a 'LoggedPage' but so it can also redirect to a 'LoggedPage' page because it's a redirection from a 'LoggedPage'
-
We don't want to try to log in on spaces where nuser is mandatory and not provided.
-
We firstly check if the xpath are present for recipient add and we also check if all recipients are internal. In other case we raise an error to investigate for a new xpath. Closes: 1063779@redmine
-
-
-