[Kst] [Bug 116461] getdata borks when the first field in the format file isn't (ASCII) alphabetically the first field.

[Ted Kisner]

On Tuesday 15 November 2005 18:50, Barth Netterfield wrote:
| It is currently conceptually depricated in kst in real time/read to end
| mode. Kst needs to know how many (complete) frames are available in a data
| source, in which it is assumed that all fields have the same size.  The
| concept of the virtual 'INDEX' field also sort of demands this.  This is
| the reason that I never brought in your getdata (you added the extra
| parameter, which I couldn't use).  It would take a pretty serious re-work
| of the data source model to change this, I think.

ok, thanks for the explanation- now I understand why my method would not work 
for this.  As George put it- the inherent flaw is that we don't have any way 
to guarantee atomic writes on a per-frame basis (except to decide which field 
gets written last, and check that field).

| The point Don was making is that defile, the real time dirfile writer,
| needs to make sure that kst know how many frames are available in *all* of
| the fields, otherwise you get weird effects (eg, X field momentairily has
| more frames than Y).

ok, so basically dirfile is meant for real-time data acquisition, and not 
general purpose binary data storage.  Or alternatively, if I have a bunch of 
different-length data files, I should make them into separate dirfiles...

-Ted