[Kst] Fwd: 2.0.7

[Nicolas Brisset]

Oh, and I forgot. But you once wrote that plugins from scripting would definitely be for 2.0.7. I have no idea how much effort it would be, but that'd be a welcome move.
I think histograms and equations have already been added, but not plugins as far as I know.

Nicolas 

----- Mail transféré -----
De: "Nicolas Brisset" <nicolas.brisset at free.fr>
À: kst at kde.org
Envoyé: Dimanche 4 Novembre 2012 00:18:13
Objet: Re: [Kst] 2.0.7

Hi Barth,

> Now that my 5 year rolling grant application has finally been
> submitted...
OK, that'll be why you've been unusually quiet the last days...
Preparing your next north/south pole "holidays" :-) ?
 
> What are the current show stoppers for 2.0.7 (or can we just
> release)?
I wanted to do some more testing before releasing, and I have not yet come around to it.
I'm not aware of any regressions right now, the last ones I mentioned have been fixed by Peter already. But some systematic testing could avoid annoying bugs as we had in 2.0.6.

Apart from regression testing, I think we should finalize some of the "time" stuff, particularly:
- in the ASCII source, adding a possibility of having fixed-sampling indexes computed as time (the same as INDEX currently, with a GUI option to indicate the sampling time)
- in the ASCII source, adding an offset either as a format string + value, or taken from the file's time/date. This would apply to all time vector options
- after testing the current svn a bit I think a separate time and date vectors as I had planned originally is not worth the effort. Peter's current implementation should be good enough
- deciding for the ASCII file buffer stuff whether the options belong in the properties of the ASCII reader or in global settings
- when selecting a time vector in the datawizard, we should automatically create plots with the X axis set correspondingly. The way it is right now, you have to create the plots and then change the x-axis settings again, which apparently confused even you. I think it would be too bad to spoil such a nifty feature in that way. Having an option at the time you read in the data and a different one to set how the x axis is displayed is OK, but we should try hard(er) to guess it right. If this point is not clear from the discussion, just play around with it and you'll understand what I mean

Regarding the points above, I have been thinking about the first 2 points and I more or less see how to do them. It should be fairly easy now that Peter has done all the heavy lifting. I am less clear about the data wizard part. So if you can/want to implement something, I'd suggest you start with the last points. Then we'll see who gets to adding the small options first :-) In any case, I think we should take the extra few days it takes to finalize that time stuff instead of releasing a half-baked implementation.

Nicolas