Made time-error message clearer

[choksi81]

On passing disallowed port numbers, time_interface was displaying misleading error in the form of raise TimeError("time_updatetime called before time_register_method!")
Also, there was an unreachable part of code raise TimeError('Error(s) in time_updatetime: ' + str(exception_list)), which has been fixed in this PR.

[vladimir-v-diaz]

@choksi81 Does this PR fix the issue you were having with the TCP relay unit test? Was it simply an invalid port number argument?

[choksi81]

As it turns out, yes, that was the problem I was facing with TCP relay
unit-tests.

On 19 November 2014 13:35, Vladimir Diaz notifications@github.com wrote:

@choksi81 https://github.com/choksi81 Does this PR fix the issue you
were having with the TCP relay unit test? Was it simply an invalid port
number argument?

—
Reply to this email directly or view it on GitHub
#154 (comment)
.

[aaaaalbert]

I ran the time unit tests against your patch, and they do pass. If any of the update_functions really raise errors, the TimeError plus exceptions_list is meaningful.

However, calling updatetime before time_register_method now results in this error:

TimeError('Error(s) in time_updatetime: []',)

(The for loop is not executed due to the empty TIME_IMP_DICT, and we directly jump into its else clause.)

We should thus raise an error before we attempt calling update_function if the TIME_IMP_DICT is empty to begin with, i.e. before the for loop.

  if not TIME_IMP_DICT:
    raise TimeError("time_updatetime called with an empty TIME_IMP_DICT. Use time_register_method to populate it!")

[vladimir-v-diaz]

9767003 now raises an exception if TIME_IMP_DICT is an empty dictionary. This PR LGTM.

@aaaaalbert May I merge some of these pull requests after reviewing them?

[aaaaalbert]

Yes, please go ahead (assuming that the unit tests don't break).