[Kst] The stand-alone GetData and kst

[D. V. Wiebe]

On Fri, Oct 24, 2008 at 10:57:57PM -0400, Barth Netterfield wrote:
> On Friday 24 October 2008 17:50:10 D. V. Wiebe wrote:
> > How about this (this is basically A1c, with an extra configure check):
> >
> > 1. Minimally fix the buggy code in /datasources/dirfile/getdata.c
> 
> Or not....

At the very minimum, I'd like to add a notice in getdata.c to the effect
that its implementation of PutData is known to be buggy and may very well
corrupt your database.  I don't think kst uses PutData, but i'd like to
to discourage users from ripping getdata.c out of their kst source tarball
to include in whatever dirfile-enabled widget they happen to be building.

> > 2. At configure time, if an external getdata library is detected, use
> > the exteral library, otherwise, use this getdata.c shipped with kst.
> 
> This is easy to do.  Basically make a new dirfile data source which uses the 
> new library (A1a).  Configure decides which one to use.

I think it's easier than that.  Since the new GetData's legacy API is backwards
compatible, all that would need to be done to use the new library would be to

a) not compile getdata.o
b) instead of including getdata.o, include the new library's .so at link time

The new GetData's legacy API should be sufficient for kst 1.x.  (i.e. the
functionality it lacks wouldn't be used by kst 1.x, anyways.)

> > This has the advantage of being fully functional for users of the
> > stand-alone GetData, while permitting Planck to keep their status quo,
> > by simply not installing the new library.
> 
> Worry.... now we have two dirfile data sources which might have bugs.
> But... there have been few bugs in the old getdata... just missing features.

The new GetData will certainly be the buggier of the two, since I've had
to re-implement parts of it to get it to do all the stuff it needs to
do.  IIRC, bugs in kst 1.x's getdata.c (at least the ones discovered by
the new GetData's test suite) are solely limited to its implementation
of PutData.

I'll try to find some time next week to implement this, if there are no
further objections.

-dvw
-- 
Don Wiebe                                   dwiebe at physics.utoronto.ca
Dept. of Physics/Univ. of Toronto
60 St. George St.                           Tele: +1-416-946-0946
Toronto ON                                  Fax:  +1-416-946-5636
Canada   M5S 1A7                            http://ketiltrout.net/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://mail.kde.org/pipermail/kst/attachments/20081105/2d50d069/attachment.sig