Re: Reduces algorithm time by prefinding footnoted grobs (issue4213042)

[Neil Puttock]

On 28 February 2011 00:32,  <address@hidden> wrote:

> For now, I'm just gonna remove this kludge and leave it as is.  It means
> that certain spanners may budge when annotations are attached to them,
> but I'll need more time to figure out a sustainable way to get the axis
> engraver to ignore the height of these annotations.  My initial hunch
> after reading your comment was to set cross-staff to #t, but this
> segfaults on any spanner that has a line break.

Why can't you just set Y-extent to #f (like BalloonTextItem does)?

> This segfault for cross staff may reveal some problems with the way that
> the FootnoteSpanner is being treated in the axis engraver - any
> thoughts?

Can you post an example which segfaults?

> I think I do need it - a pretty print to the command line shows that the
> parent of these spanners is a PaperColumn when the print function is
> called in spite of the fact that I set it to be a spanner, whereas
> "parent-spanner" doesn't change.  Somewhere in the code, this guy's
> parents are getting reset.

I can't say I've noticed this.  Every example I tested had the correct Y-parent.

> http://codereview.appspot.com/4213042/diff/34032/lily/footnote-engraver.cc#newcode119
> lily/footnote-engraver.cc:119: "Footnote ",
> On 2011/02/27 22:42:24, Neil Puttock wrote:
>>
>> "Footnote "
>> "FootnoteSpanner ",
>
>> Perhaps rename Footnote FootnoteItem.
>
> I don't mind renaming it, but could you give a reason for doing so?

It makes the distinction between the two types clearer (and ties in
with BalloonTextItem/ParenthesesItem).

> http://codereview.appspot.com/4213042/diff/34032/scm/define-markup-commands.scm
> File scm/define-markup-commands.scm (right):
>
> http://codereview.appspot.com/4213042/diff/34032/scm/define-markup-commands.scm#newcode142
> scm/define-markup-commands.scm:142: (define-markup-command (draw-hline
> layout props)
> On 2011/02/27 22:42:24, Neil Puttock wrote:
>>
>> still needs simplifying via draw-line
>
> I'm still not quite sure how to do this - do you want me to change the
> thickness property of props and then call draw-line?

See my comment on your sixth patchset (it comes with an example implementation).

Cheers,
Neil