[Kde-pim] Imcompatible behaviour change in kmime / bugs in kmime

Thomas McGuire mcguire at kde.org
Sun Sep 20 21:12:18 BST 2009


Hi,

On Sunday 20 September 2009 15:06:16 you wrote:
> The commit 1010014 breaks sending email and news article with knode,
>  because it changes the behaviour of Kmime::Message and its subclasses.
> Now the parse() and assemble() methods do not take into account
> reimplementation of virtual methods that access specials headers (like
> newsgroup(), to(), etc.).
> 
> Is it an unwanted changes that I could revert, or will stay as it is now ?

This commit was part of a review request, see 
http://reviewboard.kde.org/r/1151/. Seems like nobody saw the regression 
coming that you mention here. 
Breaking KNode was of course unwanted, shipping a KNode that can't send with 
KDE 4.4 is not an option.

As you can see in the review request, the commit fixed a bunch of things that 
were broken regarding headers. This is needed for the upcoming port of KMail's 
composer to KMime.
So the commit as a whole should not be reverted, I'd rather see a fix for this 
particular regression.
If nobody fixes this in time, we have to temporarily revert it for KDE 4.4 
though.

> This brings a few other questions about the status of kmime:
>  - when new bugs are reported and appears to be in kmime, should a test
>  case be added to the unit testing suite (waiting for someone to fix the
>  issue).

Yes, a unit test should always be added (even for new features). However, if 
the bug is not fixed, it should be marked with Q_EXPECT_FAIL, as otherwise the 
unit tests wouldn't pass anymore, and unit tests that are not passing 100% are 
not useful to detect regressions.
"Waiting for someone to fix the issue" can be problematic though, see the next 
question.

>  - who feels responsible to fix bugs in kmime 

I guess no one really feels responsible for bugs there, and everybody just 
commits there to fix bugs that are affecting himself.
But I think if someone introduces regressions he should fix them.
Constantin, are you around and available to fix this problem?
If not I guess you have to fix it yourself.

>  - what to do when unit test passes even if they are wrong.

Is there such a wrong unit test in KMime? If so, then it should be changed, 
but I wonder how an incorrect unit tests could make it into KMime.

Regards,
Thomas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
URL: <http://mail.kde.org/pipermail/kde-pim/attachments/20090920/1b5d64e0/attachment.sig>
-------------- next part --------------
_______________________________________________
KDE PIM mailing list kde-pim at kde.org
https://mail.kde.org/mailman/listinfo/kde-pim
KDE PIM home page at http://pim.kde.org/


More information about the kde-pim mailing list