lmi
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lmi] [lmi-commits] master 802eb23 6/6: Refactor for simplicity


From: Vadim Zeitlin
Subject: Re: [lmi] [lmi-commits] master 802eb23 6/6: Refactor for simplicity
Date: Mon, 5 Dec 2016 14:00:18 +0100

On Mon,  5 Dec 2016 06:55:45 +0000 (UTC) Greg Chicares <address@hidden> wrote:

GC> branch: master
GC> commit 802eb23c6b66a432a1d706ce61f7bf1d50e50e4d
GC> Author: Gregory W. Chicares <address@hidden>
GC> Commit: Gregory W. Chicares <address@hidden>
GC> 
GC>     Refactor for simplicity
GC>     
GC>     This unit test formerly reported
GC>       1608 tests succeeded
GC>     but instead now reports
GC>       650 tests succeeded
GC>     because the replaced file comparison called unit-test macros liberally.

 Hello again,

 In practice, it was very useful to see the offset of the first difference
and the bytes value immediately when running this test, this allowed me to
find one or two bugs that I had in the table binary IO code much more
quickly than if I had to compare the files manually after running the test.

 Of course, I can't argue that the new code is simpler, but mostly because
it does less. And personally I regret the part which is not done any
longer. Are you really sure it's worth doing this? After all, the old code
wasn't really complex neither (at least IMHO, it was pretty
straightforward, just longer) and I don't think simplicity or brevity of
the unit test should count for as much as that of the code of the main
program.

 Regards,
VZ


reply via email to

[Prev in Thread] Current Thread [Next in Thread]