I agree with all your comments (replied before your last message).
JB didn't speak of introducing anything new before stabilizing the timeline
refactoring.
The target is then 16.08 (1 month to go) ; if this can't be achieved, we will
have to point the release to something before the great changes (juggling with
branches).
Le jeudi 23 juin 2016, 10:17:35 CEST Harald Albrecht a écrit :
Vincent, hopefully you'll notice that I understand and highly
appreciate what JB is doing! So this is no bashing. My experience is that it
is difficult to iron out all the issue when *not* working on real
projects. Yes, this is inherently dangerous. I know that, and I'm willing to
take some pain for the sake of improving Kdenlive. I'm only asking for
starting a stabilization phase as the timeline code seems to be in a
critical situation at this time. We have really good new stuff, such as the
advanced editing and preview rendering. But if it doesn't sit on a solid
timeline code and we see corruption than we all will be in for a very bad
surprise as 16.08 (not 16.10, I'm learning *snicker*) is rapidly
approaching. Best regards, Harald
Am 23.06.2016 um 10:12 schrieb Vincent Pinon:
The timeline is handled almost in one gigantic class, with several gigantic
functions...
Several attempts to refactor this code have already been aborted.
To allow introduction of the new timeline tools, JB had no choice but to start
to untangle this spaghetti plate... and even if he fixes all the issues his
quick tests are revealing, there are side effects everywhere. Be sure he is
really concerned by all the bad experiences it causes to users.
I wish I could help him in this boring boring marathon, but really can't for
the moment.
Please support him so that he doesn't loose momentum, report him all the
details you notice (to save him from testing task), and use 16.04 for real
projects.
Vincent
Le jeudi 23 juin 2016, 09:03:28 CEST Evert Vorster a écrit :
Oh, I know _exactly_ how that feels.
Even saving often is no help, as the project looks ok, it's only when you load
it back up that the corruption becomes apparent.
I have changed my routine to be: save, then revert to last saved version. This
generally allows me to catch these timeline corruptions early.
There are a few of my bugs open on some of the timeline corruptions that I
could pin down.
Now, this is only my opinion, but it "feels" like this part of the project is
getting so complicated that small changes to the way the timeline operates has
many unintended side effects. I've seen JBM make some refactoring attempts
there, from the git commits.
Kind regards,
Evert
On 23 June 2016 at 08:52, Harald Albrecht <[email protected][1]> wrote:
The git master has some terrible issues with moving clips. I've just lost half
an hour work; have backups though. Nevertheless it is no fun to notice severe
timeline corruption only quite some time later. I've already reported several
timeline corruption issues in the past weeks.Does anybody know here when we
will enter the phase of stabilizing git master?Best
regards,Harald_______________________________________________kdenlive mailing
[email protected][2]https://mail.kde.org/mailman/listinfo/kdenlive[3]
--------
[1] mailto:[email protected]
[2] mailto:[email protected]
[3] https://mail.kde.org/mailman/listinfo/kdenlive
_______________________________________________
kdenlive mailing list
[email protected]
https://mail.kde.org/mailman/listinfo/kdenlive
