Re: *** GMX Spamverdacht *** Aw: Re: Proposal to introduce JUnit 5 in commons-numbers

[Heinrich Bohne <he...@gmx.at>]

So, if the migration from JUnit 4 to JUnit 5 has to happen
instantaneously, how about creating a new branch for the migration? It's
a lot of files that have to be refactored, which is going to take some
time, and since the changes will need to be reviewed, it would probably
be easier if the review can already take place while the migration is in
progress, rather than all at once when it is finished. Also, it would
reduce the risk of conflicting pull requests, thus potentially saving time.

On 5/27/19 8:28 AM, Sven Rathgeber wrote:
> +1
>
> We did the migration 4 -> 5 for various products at work. It was straightforward.
> I vote for a clean cut, as well.
>
> s/Before/BeforeEach/
> s/After/AfterEach
> .... and you are halfway through.
>
> Sven
>
> Gesendet: Montag, 27. Mai 2019 um 07:28 Uhr
> Von: "Eitan Adler" <li...@eitanadler.com>
> An: "Heinrich Bohne" <he...@gmx.at>
> Cc: "Commons Developers List" <de...@commons.apache.org>
> Betreff: Re: Proposal to introduce JUnit 5 in commons-numbers
> (please make sure to CC me directly on replies)
>
> On Sun, 26 May 2019 at 17:17, Heinrich Bohne <he...@gmx.at> wrote:
>
>> ... (I didn't subscribe to the list, which I now regret as I am editing
>> the header fields of this email manually, hoping that I do it correctly and
>> that the message will be attributed to the correct thread).
>>
> This is the first time I seeing this thread.
>
>
>> Regarding the point about a potential mess of multiple JUnit versions:
>> JUnit 5 seems to provide a "Vintage" test engine to support the execution
>> of JUnit 3 and 4 tests alongside JUnit 5 tests.
>>
> It is precisely this that I am advocating against. While JUnit 5 has a
> method of running older tests, by leaving the older style tests around
> we're producing confusion as to how it will get run or how to write new
> tests. Using it as a migration tool is reasonable, but the commons are
> small enough that wholesale migration is "easy".
>
>
>>
> --
> Eitan Adler
>
>
>
>


---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org

Re: *** GMX Spamverdacht *** Aw: Re: Proposal to introduce JUnit 5 in commons-numbers

[Eitan Adler <li...@eitanadler.com>]

On Tue, 4 Jun 2019 at 16:09, Heinrich Bohne <he...@gmx.at> wrote:

> So, if the migration from JUnit 4 to JUnit 5 has to happen
> instantaneously, how about creating a new branch for the migration? It's
> a lot of files that have to be refactored, which is going to take some
> time, and since the changes will need to be reviewed, it would probably
> be easier if the review can already take place while the migration is in
> progress, rather than all at once when it is finished. Also, it would
> reduce the risk of conflicting pull requests, thus potentially saving time.
>

I don't think it has to be done instantaneously but it shouldn't drag on,
or be done "as we go". In terms of creating a branch: seems reasonable to
me. I'll happily contribute and review.

That said, the first change we need to make is to modify `pom.xml` to
support running JUnit 5 tests + the vintage runner. I don't know maven too
well but took a stab at it here:
https://github.com/apache/commons-numbers/pull/47


-- 
Eitan Adler