Re: Cleanup list of patches and bug reports on sf.net

[Harald H. <h.h...@tu...>]

On Sat, 23 Oct 2004, Ethan Merritt wrote:

> On Saturday 23 October 2004 01:55 pm, Harald Harders wrote:
> > I have seen some old patches on sf.net that may have to change the stat=
us.
> >
> > For example, the patch #982765, "Patch for AI (Adobe Illustrator) term"=
,
> > has been discussed here. Since the whole AI terminal is outdated and
> > postscript is understood by Adobe Illustrator, this patch could be
> > rejected and closed.
> > Or patch #743667, "Epslatex term merged w/ pslatex/pstex", is superseed=
ed
> > by #1040192, "Merge post, epslatex, pslatex, pstex terminals", and coul=
d
> > also be closed.
>
> It would be nice to hear from people actually using these various
> drivers.  I suppose we can deprecate ai.trm and the older forms of
> the various latex terminals, but we may have to resurrect them if
> there are user communities who rely on some feature of the older
> terminal drivers that we didn't realize we were changing.

When closed, patches are not deleted immediately, right? Then, this should
not be a too big problem.

> > And isn't the bug report #963176, "wish: only create docs for available
> > terminals", solved?
>
> Not that I know of.

I think the discussion was somehow good. In all platform-independent
formats, all terminals should be described. In the online form (via typing
help in gnuplot), either only installed ones should be described (which is
the case, I think) or not installed terminals should be described but
marked as not installed. But the bug report #963176 requested to not
describe uninstalled terminals at all, if I remember correctly. And then,
it is not unsolved.

> > This patch fixes bug #1000676 =A0(Rotated multiline text misaligned).
> > The alignment of rotated multiline text was correct for
> > horizontal and vertical text but incorrect for other
> > angles (they were handled as not rotated). This patch
> > corrects this.
>
> I don't agree that this was a bug.  It depends on what you want to
> do with the text.  For me, the chief use is to squeeze more (or longer)
> labels along an axis than would otherwise fit.  And in this case the
> baseline should remain the axis itself, i.e. not rotated. IMHO it looks
> really odd to have multi-line axis tic labels swing out and away from
> the axis.

That is of course a usage I have not thought of, yet. But it surely is a
bug to have the old alignment when using a rotation angle of more than 45
degree. Then, the lines overlap in most cases. But it should be possible
to rotate the whole text block, too.

> If you think it is useful to have the entire text block rotate as a
> unit, I think we would need some new syntax or justification
> option to specify it.  Can you give an example of when you
> would want this?

I have had some plots with legends that had up to 2 lines each and were
rotated by 60 degree. I wanted them to be centred. And it looked very odd
that they were centred to a vertical(!) line. To me, it were best to
centre to a line rotated by 60 degree from the vertical direction.

In short: I still think it is a bug since even text rotated by 89 degree
is aligned vertically. If the alignment axis should rotate with the text
or if it has to stay at a fixed angle up to a threshold angle is a
question of taste. I think both possibilities should be there.
Maybe there should be an additional keyword 'alignmentangle' or similar.

Greetings
Harald

--=20
Harald Harders
h.h...@tu...
http://www.harald-harders.de