Martin von Zweigbergk <martinvonz@google.com> [Fri, 20 Nov 2020 10:34:26 -0800] rev 45891
errors: remove ParseError.args
With the previous few patches, it is no longer needed (as far as the
test suite can tell anyway).
Differential Revision: https://phab.mercurial-scm.org/D9354
Martin von Zweigbergk <martinvonz@google.com> [Fri, 20 Nov 2020 13:55:32 -0800] rev 45890
errors: let ParseError use Abort.__bytes__
The function is no longer used anywhere as far as our tests can tell,
so there's no need to have a custom version of it.
Differential Revision: https://phab.mercurial-scm.org/D9353
Martin von Zweigbergk <martinvonz@google.com> [Fri, 20 Nov 2020 10:31:56 -0800] rev 45889
config: clean up message about ignored untrusted config
The error message relied on Python's default formatting of arguments
to an Exception's constructor. Let's try to make it a little more
readable for users.
Differential Revision: https://phab.mercurial-scm.org/D9352
Martin von Zweigbergk <martinvonz@google.com> [Fri, 20 Nov 2020 10:22:58 -0800] rev 45888
tests: use new ParseError.format() in test-trusted.py
Differential Revision: https://phab.mercurial-scm.org/D9351
Martin von Zweigbergk <martinvonz@google.com> [Thu, 19 Nov 2020 15:13:39 -0800] rev 45887
errors: make ParseError a subtype of Abort
I didn't plan this before, but the previous two changes made it really
easy to make `ParseError` a subtype of `Abort`. It seems obvious with
hindsight that I *should* have planned it :)
Differential Revision: https://phab.mercurial-scm.org/D9350
Martin von Zweigbergk <martinvonz@google.com> [Fri, 20 Nov 2020 13:24:45 -0800] rev 45886
tests: make doctests not depend on str(ParseError()) format
`ParseError` implements `__bytes__`, but it doesn't implement
`__str__`, so it gets the default `__str__` implementation. The next
patch will make it so `ParseError` gets a `__str__` implementation,
which changes the format slightly. This prepares by making us not
depend on the format.
Differential Revision: https://phab.mercurial-scm.org/D9349
Martin von Zweigbergk <martinvonz@google.com> [Fri, 20 Nov 2020 09:17:38 -0800] rev 45885
errors: format "abort: " text in a new Abort.format() method
This remove some duplication we had.
Differential Revision: https://phab.mercurial-scm.org/D9348
Martin von Zweigbergk <martinvonz@google.com> [Fri, 20 Nov 2020 08:51:45 -0800] rev 45884
errors: make formatparse() an instance method on ParseError
It's just a little simpler this way.
Don't ask me what the "hg: " prefix signifies to the user, I just left
it as it was. I think we should consider changing the prefixes later
(maybe always use "abort: ", or maybe use more specific prefixes in
general).
Differential Revision: https://phab.mercurial-scm.org/D9347