Re: [Plplot-devel] Wish list for the release--various thoughts.

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

On Fri, 18 Jan 2002, [iso-8859-1] Jo=E3o Cardoso wrote:

> On Thursday 17 January 2002 17:15, Alan W. Irwin wrote:
> | > When will be 5.1 released?
> |
> | 24 January.
>
> !?!
>
> I thought that this release would be 5.0.5! But if I were the boss we
> would continue until reaching release 5.0.31415926

;-)

>
> Seriously, there are too many new functionalities for this to be 5.1.
> Java is all new, python has suffered significative modifications,

I don't care that much about version numbers except that I don't want to go
through them too rapidly.  My understanding is the major number should be
incremented for huge changes, the minor number should be incremented for
fairly large changes, and the patch number incremented for smaller changes.
I absolutely agree with you there have been some fairly large changes made
since 5.0.4.  Thus, I assumed it was appropriate to increment the minor
number and move on to 5.1.0.  But if the developers here feel that 5.0.5 is
more appropriate, that is certainly fine with me!

For the others here, please let me know your feelings on this.

> [...] what else?... I can't use "plrender"
> with device tk,..

Neither can I, but this is the first I have heard of it!  (I tend to
do strictly non-interactive tests, and leave the interactive stuff to
the rest of you.)  I must have missed your bug report to the list....;-)

Here is mine....

Maurice, here are the symptoms (with dynamic drivers configured if that
makes a difference)

There is no problem if I executed

=2E/x01c -dev tk

However,

=2E/x01c -dev plmeta -o x01c.meta
=2E/plrender -i x01c.meta -dev tk
TCL command "
         invalid command name "

Can this bug be fixed before the release?

>
> [....] My not yet commited plimage() changes and related
> functions are yet alpha. At this point I'm afraid of commiting my
> changes, as perhaps the current cvs plimage() is more stable than
> mine.

But the point is this does not affect anything other than x20c.  It would b=
e
a much tougher decision if you were fiddling with the very inner core of
plplot, but you are not.  I hope to get out the next release within a month
or two so if you make a mistake with plimage now, it will soon get
corrected.  So I am inclined to say that if x20c still works on your
platform with all your uncommitted changes, then go ahead and commit them.
However, if you are not comfortable with that, that is fine as well.
Ultimately, of course, it is your decision.

> I have just added an option to x01c to show how one can
> programatically save a file (that is stable, I will commit it).
>
> What I mean is that 5.1 is too definitive for the current status.
> As you say, we are not many, and can't afford the pressure of a
> faulty major release.

My understanding of release traditions is quite different.  Patch version 0
of anything is never expected to be reliable.  Compare kernel 2.4.0 with
2.4.17 or KDE 2.2.0 with 2.2.2, for example.

That said, I do want the functionality that worked in 5.0.4 to continue to
work in 5.0.5 (or 5.1.0., whichever the developers here prefer).  I believe
that is the case.  It is only a small subset of the new things that have
been added that are not completely stabilized yet.  For example, dynamic
drivers seem to work perfectly for all but xwin and tk, and our
configuration scheme works around that problem so no user will be affected
by it (assuming that plrender bug gets fixed.) Java is of course considered
to be experimental since it is still incomplete, but for the defined API, i=
t
is rock solid.  Python no longer segfaults.  Thus, I will have no
compunction about telling all our users we no longer support 5.0.4, and it
is time to move to this next release (whatever version number we decide).

Alan