Testing reports

[Colin Law]

On 12 April 2012 15:06, John Layman <john.layman at laymanandlayman.com> wrote:
>> Tests which fail for insignificant reasons (like case in an SGML tag)
> waste
>> developer time and are worse than useless.
>
> One could argue that it's a bigger waste of developer time to complicate
> test creation in the interest of discriminating that which is significant
> from that which is not in every imaginable potential scenario.  By
> implication, you appear to be saying that the trivial work of updating a
> test oracle is a poor use of the programming artiste's time.  Artists do
> clean their own brushes, you know.  Doing so may be trivial, but it's part
> of the process.

I think the point is that with the right tools it is easier to code a
test that says that there should be a div of class X containing
paragraph with id Y with content Z than it is to test for a complex
string of html.  Therefore it /should/ be easier just to test what is
important.  If however the tools are not available then it may be
necessary to go for a simple exact match for the whole page and
somehow work around the fact that this may cause complications with
different build environments for example and is more work if the
report is updated.

Colin

>
> -----Original Message-----
> From: gnucash-user-bounces+john.layman=laymanandlayman.com at gnucash.org
> [mailto:gnucash-user-bounces+john.layman=laymanandlayman.com at gnucash.org] On
> Behalf Of John Ralls
> Sent: Wednesday, April 11, 2012 11:07 PM
> To: gnucash at double-bars.net
> Cc: warlord at MIT.EDU; gnucash-user at gnucash.org
> Subject: Re: Testing reports
>
>
> On Apr 11, 2012, at 5:06 AM, Colin Scott wrote:
>
>>
>>> I would argue that the contents of the report are more important than
>>> the formatting, at least in the context of how this thread started.
>>
>> Hmmmmm.  I'm not at all sure that I agree.  *WHY* would the report
> formatting change?
>>
>> The fundamental point here is that even if the change is harmless it
> should only occur as the result of a specific and specifically intended
> action - anything else must be a bug!!!  Either way, any change in the
> output should be flagged, either so it may be certified as correct (and then
> incorporated into the standard against which subseqent tests will be made)
> or so the cause of the change can be tracked down and fixed.
>>
>> There are no circumstances I can envisage where it is proper to ignore
> *any* change to the output.
>
> Rubbish. Non-trivial programs have both interdependencies between modules
> and dependencies on external libraries. Those can -- and do -- change, and
> can cause changes in the behavior or output of a module under test. Well
> designed and written tests focus developer effort on changes which matter
> and ignore changes that don't. Tests which fail for insignificant reasons
> (like case in an SGML tag) waste developer time and are worse than useless.
>
> Regards,
> John Ralls
>
>
> _______________________________________________
> gnucash-user mailing list
> gnucash-user at gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-user
> -----
> Please remember to CC this list on all your replies.
> You can do this by using Reply-To-List or Reply-All.
>
> _______________________________________________
> gnucash-user mailing list
> gnucash-user at gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-user
> -----
> Please remember to CC this list on all your replies.
> You can do this by using Reply-To-List or Reply-All.



-- 
gplus.to/clanlaw