[kdepim-users] kde-pim hopeless?
Martin Steigerwald
Martin at lichtvoll.de
Mon May 6 20:24:31 BST 2013
Am Sonntag, 5. Mai 2013, 08:04:02 schrieb Russ Kepler:
> Should I just give up and move on? From my perspective the kontact/kmail
Well that is entirely your decision.
> system has gone from a fairly robust (if occasionally slow) system that
> dealt well with mail to a bug-ridden system that doesn't filter received
> mail properly, search received mail properly after it's filtered and sucks
> down resources like nothing I've ever seen (virtuoso-t usually is at 2GB of
> ram after a couple of days and is usually consuming one core continuously).
I agree that even in 4.10.2 it still has a lot of issues. I reported more than
twenty bugs in about two weeks.
I see huge issues with filtering, race conditions with mail moving and
deletion, a data loss bug with my CRM114 filter rules, huge resource usage in
possible default setup for POP3 accounts and even some issues I didn´t yet
report, like all mails disappearing from a message list view if I press Ctrl+J
on *one* mail in it - luckily it appears after selecting a different folder and
going back.
I think KDEPIM still needs lots of stability and robustness and correctness
work starting from any bugs that may cause data loss. I think it also needs
testing for data safety on moving, copying mails, deleting mails, filtering
mails tons of mails, also when being put under load.
> For a while it looked like things were getting better then suddenly it all
> regressed to worse than the starting state. I don't know if the development
> team does any unit testing but I'm fairly sure that there's no regression
> testing, or if there is the testing is either too minor, test cases too
> trivial or the results of testing ignored.
>
> It's PITA to move away but unless someone can give me any hope that things
> are going to get better I'm feeling that's the only choice. I'd rather
> have a system that ran slowly than one I can't trust at all.
I can understand your thoughts. I thought along those lines as well.
While I found a way I can use KDEPIM-2 4.10.2 with some restriction like
letting mails flowing into a special inbox to weed out spam manually and Ctrl+J
filter the safely it seems. I wouldn´t trust it to move 10000 mails from one
folder to another and things like that.
But frankly, I just don´t want to leave it. I used KDEPIM since August 2003, I
just don´t want to move over all my PIM data, especially not the mail to
something else. I like the GUI, I like the features, I like lots of little
details.
I don´t think that the developers do not want to improve KDEPIM-2. They are
trying hard. I do think its more a lack of time.
Thus I am prepared to live with some limitations like described above for a
while. And I am willing to invest some time to isolate test cases that may
lead to easier reproducability of issues. My current idea is to simply my
reproducer for the mail loss issues I found with my CRM114 spam filter rules to
using a separate user with those 5000 or 10000 script generated mails and use
a bog standard configuration with just those 5 filter rules and an empty maildir
from scratch.
When I have that test setup, I can move around mails, copy them, copy folder
structures and see what works.
For me its also an issue with time. Testing and reducing bugs to easily
reproducable isolated test cases seems to be challenging to me with such a
complex software - necessarily complex at times. I would love to work with
other users on this. Maybe share workload. Maybe a bug triaging day? However
in my current setup, I think due to space constraints on my SSD its not easy
to afford compiling KDEPIM from source.
But I think that is where we users can help. Trying to get to the bottom of
bugs and find "to reproduce this, you do this with an empty user". Still even
then a dev needs to have time to look into this.
I don´t know an easy answer or solution, but I think it still needs lots of
work.
I appreciate that Laurent Montel and others do actually work on fixing bugs a
lot these days, from what I can see in KDE Commit Digests, but I also fully
understand that with those 20 bug reports I already added quite lot work
already.
KDEPIM-1 at some time also had quite nasty bugs with index files, "no subject"
mails and other stuff, that took devs quite some time to fix as well. But I
fully understand, if a user in the current state of affairs does not trust it
too much. I don´t trust it very much either. I leave 30 days of mails on the
POP3 server for a reason and am very reluctant to enable automatic filtering
which gave my move mails race conditions for example. And I have no motivation
at all to enable the CRM114 spam filter rules, until I am confident that the
one-byte sized file bug has been fixed for good. I also do not trust it too much
100000 mails from one place to another and things like that. I did trust
KMail-1 with that, although I made it in smaller pieces, cause KMail-1 would
be blocked for ages otherwise. But whats important: It did complete the work
and the result was proper.
Ciao,
--
Martin 'Helios' Steigerwald - http://www.Lichtvoll.de
GPG: 03B0 0D6C 0040 0710 4AFA B82F 991B EAAC A599 84C7
_______________________________________________
KDE PIM users mailing list
Subscription management: https://mail.kde.org/mailman/listinfo/kdepim-users
More information about the kdepim-users
mailing list