>>>>> John Arbash Meinel <email address hidden> writes:
> ...
>> The existing daughter classes already either define a specific _fmt and/or
>> don't accept orig_error as a parameter so aren't concerned by this fix.
>>
>> > For now, though, it seems fine. We usually add tests into
>> > bzrlib/tests/test_errors.py for formatting stuff like this, though.
>>
>> For use facing errors that's true, in this specific case I'm addressing
>> debug concerns, we should not use InvalidHttpResponse for user-facing
>> errors.
> I thought the whole point of the original bug was that this got shown as
> a *user-facing* error, and that it made it confusing what the original
> failure was.
Well, it's an unexpected exception: and invalid http response, this is
not some kind of error we expect and for which we can give a better
explanation to the user, it's used as a catch-all for several kind of
low-level errors.
>>>>> John Arbash Meinel <email address hidden> writes:
> ...
>> The existing daughter classes already either define a specific _fmt and/or tests/test_ errors. py for formatting stuff like this, though.
>> don't accept orig_error as a parameter so aren't concerned by this fix.
>>
>> > For now, though, it seems fine. We usually add tests into
>> > bzrlib/
>>
>> For use facing errors that's true, in this specific case I'm addressing
>> debug concerns, we should not use InvalidHttpResponse for user-facing
>> errors.
> I thought the whole point of the original bug was that this got shown as
> a *user-facing* error, and that it made it confusing what the original
> failure was.
Well, it's an unexpected exception: and invalid http response, this is
not some kind of error we expect and for which we can give a better
explanation to the user, it's used as a catch-all for several kind of
low-level errors.