Re: [wtp-dev] WTP modules unit tests quality?

[<p.bodnar@xxxxxxxxxx>]

Hello,

thanks for your answers. I must say I have expected such or a similar reply :) It would really probably require a dedicated person with direct access to the repositories to perform the tests' cleanup effectively.

Ad StringCompareUtil.equalsIgnoreLineSeperator - yes, I got the idea behind this method, I take your reply as a summary for the not-that-familiar with the mentioned code.

Ad contributions - well, I would be happy if anyone reviewed and gave feedback to the proposed changes in the previously mentioned bug (probably too fresh yet). It's just one module and one test, but maybe someone (as described above) could do more with that?

Ad problematic old test classes - I have selected these rather randomly, when "surfing" the test classes from other modules, so I can't say how many "broken windows" are there among newer tests, hopefully less, as you have written.

In the meantime, before I could ask here, someone else has probably answered my question regarding the purpose of having a separate Git repository for tests - please see https://bugs.eclipse.org/bugs/show_bug.cgi?id=422915#c4 (another fresh bug). Don't you find it interesting? Do you think the same merging could be done with the "wtp.sourceediting" project?

Thanks again, enjoy the holidays :)

PB

______________________________________________________________
> Od: NICHOLAS SANDONATO <nsandona@xxxxxxxxxx>
> Komu: "General discussion of project-wide or architectural issues." <wtp-dev@xxxxxxxxxxx>
> Datum: 18.12.2013 07:24
> Předmět: Re: [wtp-dev] WTP modules unit tests quality?
>
>Hi,
>
>Thank you for raising your issues. The 
>StringCompareUtil.equalsIgnoreLineSeperator (good typo catch), is used to 
>handle potential problems with the formatted output ending up with a 
>different line-ending due to running the tests on different platforms than 
>the expected output was generated on. There may be better solutions to 
>this.
>
>Speaking for SSE's components, since these were mentioned in particular, I 
>don't think there has been any official review of style for these test 
>cases. Additionally, a lot of these test cases are very old. I think we're 
>always happier to see at least some test cases included with new 
>contributions rather than none at all. We'd certainly love to have 
>high-quality tests, and any contributions towards getting us there would 
>be appreciated. With the number of tests that we have, it would be a 
>pretty big undertaking.
>
>Thanks,
>Nick
>
>
>
>From:
><p.bodnar@xxxxxxxxxx>
>To:
><wtp-dev@xxxxxxxxxxx>
>Date:
>12/14/2013 01:10 PM
>Subject:
>[wtp-dev] WTP modules unit tests quality?
>Sent by:
>wtp-dev-bounces@xxxxxxxxxxx
>
>
>
>Hello,
>
>I have opened a new bug on issues in the WTP's XML formatter unit tests: 
>https://bugs.eclipse.org/bugs/show_bug.cgi?id=424054. As I also noticed 
>tests-related issues in the other WTP modules (projects), I decided not to 
>write it to the bug itself, but this mailing list hopefully is a better 
>target. So here it goes:
>
>I noticed that the problematic 
>StringCompareUtil.equalsIgnoreLineSeperator() (shouldn't it be sepArator?; 
>mentioned in the bug) is used in other WTP modules (jsp, css, html) as 
>well. So maybe the same problem there?
>
>Next: Test classes across and even within different modules follow 3 or 
>more naming conventions (even "*Tests" - which won't be executed, by 
>default, if you are ever about to use Maven (Tycho?), see 
>http://maven.apache.org/surefire/maven-surefire-plugin/examples/inclusion-exclusion.html
>).
>
>Another: Have somebody seen and reviewed classes like 
>DOMImplementationTests or TestAttributesOrder? I know most people say 
>"it's just test", but anybody having some practice in writing JUnit code 
>must see the problems (repeated code, setUp() trying to be called just 
>once by using non-static variable, ...). No offence, authors :)
>
>I understand that these kind of problems are harder to be solved (code 
>review? problematic on open source project?) and there are "real" bugs to 
>be fixed but we should also bear in mind the long-term benefits (incl. 
>attracting, or at least not dis-attracting, new possible contributors) 
>when solving or warn about these issues. Hopefully the "production" code 
>is treated better. What do you think?
>
>Best regards
>
>PB
>_______________________________________________
>wtp-dev mailing list
>wtp-dev@xxxxxxxxxxx
>https://dev.eclipse.org/mailman/listinfo/wtp-dev
>
>
>
>
>
>----------
>
>_______________________________________________
>wtp-dev mailing list
>wtp-dev@xxxxxxxxxxx
>https://dev.eclipse.org/mailman/listinfo/wtp-dev
>