[Plplot-devel] Is it time to remove our old fortran and our old Tcl bindings and examples?

[Alan W. I. <ir...@be...>]

Hi Arjen:

This e-mail concerns cruft removal for both the Fortran and Tcl cases.

Fortran:

We introduced the new Fortran binding and corresponding examples as of
PLplot-5.12.0 (released on 2017-01-29), but as a safety measure
supplied the old Fortran binding and examples if the user specified
the CMake option -DPL_DEPRECATED_fortran=ON (which gave access to
bindings/old_fortran and examples/old_fortran).

Since that release I think we have gotten not even one complaint about
the new fortran binding which introduced many backwards incompatible
changes.  (There was one complaint about lack of
Fortran support in CMake for a non-mainstream Fortran compiler, but
that is a different issue that affects both our old and new fortran
binding for that one user.) Furthermore, the next release is likely
coming out in early 2018 or roughly one year later than PLplot-5.12.0.
Therefore, I believe this release cycle is the right time to get rid
of the cruft consisting of the -DPL_DEPRECATED_fortran=ON option and
bindings/old_fortran and examples/old_fortran.

If you agree with this timing, and assuming nobody on the
plplot-general list objects (i.e., there is nobody there who is still
actually using the old fortran binding), then I would plan to remove
this cruft within a few days from now (just to make this change early
in the release cycle so it gets maximum testing before the release in
early 2018).

Tcl:

There is a similar story with the redacted Tcl API introduced in
PLplot-5.12.0.  There have been no complaints about that large
backwards incompatible change, and the next release will be roughly a
year away from the release where this change was introduced. 
Therefore, I believe this release cycle is the right time to get rid
of the cruft consisting of the -DUSE_NON_REDACTED_TCL_TK=ON option and
the corresponding bindings/non_redacted_tcl, bindings/non_redacted_tk,
bindings/non_redacted_tk-x-plat, examples/examples/non_redacted_tcl,
and examples/non_redacted_tk.

If you agree with this timing, and assuming nobody on the
plplot-general list objects (i.e., nobody there is using the
-DUSE_NON_REDACTED_TCL_TK=ON option), then I would plan to remove this
cruft within a few days from now just as in the fortran cruft removal
case and for similar reasons.

General remarks about this cruft removal:

I am pushing this cruft removal because keeping those old files around
implies we should sometimes test the -DPL_DEPRECATED_fortran=ON and
-DUSE_NON_REDACTED_TCL_TK=ON options (which I don't want to do) and
also do minimal maintenance to the old versions of the files (which I
also don't want to do).  For example, I am about to remove the
long-deprecated plrgb, plrgb1, plhls, and plwid.  But some of those
are assumed to be available (when -DPL_DEPRECATED=ON) in the old
versions of the Fortran and Tcl bindings.  Which implies I need to
either maintain those old versions consistent with the removal of
plrgb, plrgb1, plhls, and plwid or do the much simpler task of
deleting those old versions.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state
implementation for stellar interiors (freeeos.sf.net); the Time
Ephemerides project (timeephem.sf.net); PLplot scientific plotting
software package (plplot.sf.net); the libLASi project
(unifont.org/lasi); the Loads of Linux Links project (loll.sf.net);
and the Linux Brochure Project (lbproject.sf.net).
__________________________

Linux-powered Science
__________________________