[Kst] ASCII reader

[Peter Kümmel]

On 23.10.2012 23:12, Brisset, Nicolas wrote:
>>> UI:
>>> -The ascii config dialog does not fit on a 1366x768 screen - its too
>>> tall.  It should be made wider and shorter.
>> OK, some fiddling with designer.
> Peter has done some tweaking. I think we still have some points to discuss on that dialog, like for instance the fact that I think time/date options should go to a dedicated group plus some other details.
> I actually like doing this UI tweaking, I can finalize it when we have clarified all the implementation bits and tricks.
>
>>> -The purpose of the QDateTime format string field in the axis tab is
>>> not clear.  It doesn't seem to have anything to do with how the plot
>>> appears (everything else in the dialog does - since it is the plot
>>> dialog).  Is this a copy of the field in the ascii config dialog?
>> You could define the format of the time, e.g. suppressing the year at the
>> printed time, e.g.:
>> 17:00.01 [min]
>> by selecting "QDateTime format string" in the combobox and using the string
>> "hh:mm:ss"
> OK, now I get it. It wasn't clear to me either. The options in the ASCII config dialog tell Kst how to turn the ASCII strings into C time (or whatever the internal representation is) and is specific to ASCII. This part in the plot dialog tells Kst how to interpret/display the values and applies to all data, regardless of the file type is was read from, possibly with less options than ASCII will soon have.
> In fact, this raises 2 questions:
> 1) in the numerous options available from the "Display as" dropdown, how many are really used and how many could be superseded by the "free" formatting string Peter added? Personally, I can't even guess what most of them mean: JD, MJD, RJD: ???

Yes, some options are now redundant. But they are needed for backward compatibility.

> 2) Kst should be smart enough to pre-select the appropriate X-axis settings when time/date are configured when reading the file. This probably means checking supportsTime() from the wizard, and setting axes accordingly when creating the plots. Then the whole complexity will be hidden to the user, unless he wants to do something special.

There are several time functions in the datasource interface which are nowhere used ATM.
Maybe the idea was to use them for such stuff.
There were some calls in Kst1 to these functions but I haven't figured out I they were really used.

Peter