On Tue, Dec 28, 2010 at 12:14 AM, Ben Abbott <address@hidden> wrote:
> On Dec 26, 2010, at 4:19 AM, Konstantinos Poulios wrote:
>
>> 2010/12/24 Jordi Gutiérrez Hermoso <address@hidden>:
>>>> On Dec 22, 2010, at 6:43 AM, logari81 wrote:
>>>>
>>>>> Currently, the calculation of the extent of text objects in Octave is
>>>>> very limited. For example it doesn't take into account the rotation
>>>>> angle:
>>>>>
>>>>> Revision 11407 of Octave:
>>>
>>> This number doesn't make sense in general. Changeset numbers are local
>>> to a repository. While it looks likes people have been making a strong
>>> effort to keep the hg history linear, in my view an artficial
>>> representation of what's actually happening in development, as soon as
>>> you break linearity of the history, it becomes impossible to
>>> universally number changesets. In a distributed VCS, this break from
>>> linearity will happen sooner or later (in fact, it's already happened
>>> a few times, right now buried deep in the history).
>>>
>>
>> yes, I am aware of this property of distributed cvs types. Actually,
>> when I refer to version numbers, these do not correspond to my local
>> repository but to the official one here:
>>
http://hg.savannah.gnu.org/hgweb/octave/
>> In this case revision 11407 is:
>>
http://hg.savannah.gnu.org/hgweb/octave/rev/ed827ffa5a43
>>
>>> On 23 December 2010 20:53, Ben Abbott <address@hidden> wrote:
>>>> This changeset doesn't apply for me. In particular, some lines from
>>>> graphics.cc seem to be different than what is indicated by your
>>>> changeset.
>>>
>>> More telling, the parent node for this changeset is
>>> f94a71bf5e88151529c2ee7ecdf10c6f5ad29aa9, which doesn't exist in the
>>> Savannah repo.
>>
>> oups, sorry for this, I tried to merge three commits into one but
>> apparently I missed the first one. In the attachment you can find the
>> complete version, hopefully with no further errors (I had to recover
>> the the first part of the changes from my memory, but it seems to work
>> well).
>>
>>>
>>> Kostas, all you have to do is provide the changeset with that hash,
>>> which should exist in your local repository. The hg export command
>>> lets you specify more than one changeset, so it looks like you have
>>> more than one local commit than what you've shared.
>>>
>>> - Jordi G. H.
>>>
>>
>> Kostas
>> <text-extents.changeset>
>
> The changesets apply cleanly for me. However, for printed output the text objects no longer show up.
>
> I've attached pdfs produced using the fltk backend with the sources patched and unpatched.
>
> backend fltk
> close all
> demo text
> print -dpdfwrite test.pdf
>
> My tip is ...
>
> changeset: 11416:74e285bb61c9
> tag: tip
> user: Ben Abbott <address@hidden>
> date: Mon Dec 27 11:13:06 2010 -0500
> summary: text.m: Add demo for text rotation and alignment.
>
> Ben
Please try the attached patch. It should fix the problem with plot printing.
Actually, there are more changes necessary in order to make everything
consistent but this should be sufficient for recovering the previous
functionality.
Kostas
text-extents-1.changeset
patched-1-fltk.pdf