While generally looks as a good idea, the restriction of setters visibility
doesn't seem a valuable change in FontBox (like many other libraries)
because there are already quite a lot of public setters (e.g.
org.apache.fontbox.ttf.GlyphData#setNumberOfContours) which could allow
users to accidentally or deliberately spoil the data gathered from the font.
So perhaps it would be enough to introduce a new method only (to get a view
of the components) and avoid a breaking change.
Cheers,
Vladimir
вт, 10 янв. 2023 г. в 01:48, Tilman Hausherr <thaush...@t-online.de>:
> On 08.01.2023 13:51, Andreas Lehmkuehler wrote:
> > Am 08.01.23 um 13:28 schrieb Tilman Hausherr:
> >> Hi,
> >>
> >> The idea sounds ok to me, the only risk I see would be that somebody
> >> changes a GlyfCompositeComp element.
> > How about reducing the visibility of both setters to protected and as
> > long as we use an unmodifiable list to provide the components there
> > shouldn't be any risk or am I overlooking something?
>
> Yes this makes sense. But we can't do it for 2.0 because this would be
> an API break. (We can add a comment)
>
> Tilman
>
>
>
> >
> > Andreas
> >
> >>
> >> Tilman
> >>
> >> On 07.01.2023 14:15, Vladimir Plizga wrote:
> >>> Hello,
> >>>
> >>> I'm starting to use the FontBox library (tried both GA version
> >>> 2.0.27 and
> >>> the latest pre-release 3.0.0-alpha3) to inspect composite glyphs in
> >>> fonts.
> >>> In particular, I need to know what exact glyphs (components) are
> >>> included into a composite glyph and what transformations have been
> >>> applied
> >>> to them. However, when I obtain information about a glyph I can only
> >>> find
> >>> out whether it is composite and how many components are in it. There
> >>> is the
> >>> org.apache.fontbox.ttf.GlyfCompositeComp class that has (almost) all
> >>> the
> >>> necessary information, but there is no way to get it from the library
> >>> (AFAIS). The closest method to it is the
> >>> org.apache.fontbox.ttf.GlyfCompositeDescript#getCompositeComp, but
> >>> it is
> >>> not suitable because: (1) it's private and (2) it takes point index
> >>> as an
> >>> argument which is out of scope for this particular task (as I work
> >>> with the
> >>> glyph at whole but not with its particular points).
> >>>
> >>> I suppose there is some deliberate design decision behind this
> >>> limitation,
> >>> but it really hinders "to obtain low level information from font
> >>> files" (as
> >>> stated in the library description).
> >>>
> >>> As a possible solution, I would suggest to supplement
> >>> org.apache.fontbox.ttf.GlyfCompositeDescript class with a new method
> >>> like
> >>> this:
> >>>
> >>> /**
> >>> * Gets a view to the composite components
> >>> * @return unmodifiable list of this composite glyph's {@linkplain
> >>> GlyfCompositeComp components}
> >>> */
> >>> public List<GlyfCompositeComp> getComponentsView() {
> >>> return Collections.unmodifiableList(components);
> >>>
> >>> }
> >>> , which is aimed to provide a read-only view of the components of a
> >>> composite glyph. Or provide a similar method for getting a particular
> >>> component by its index. In either case it would also be needed to
> >>> make the
> >>> static flag masks public in org.apache.fontbox.ttf.GlyfCompositeComp
> >>> class
> >>> (since now they are protected).
> >>>
> >>> Could you please comment on the described limitation and the suggested
> >>> solution (if it makes sense of course)?
> >>>
> >>> Cheers,
> >>> Vladimir
> >>>
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org
> >> For additional commands, e-mail: users-h...@pdfbox.apache.org
> >>
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org
> > For additional commands, e-mail: users-h...@pdfbox.apache.org
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org
> For additional commands, e-mail: users-h...@pdfbox.apache.org
>
>
