Object clipped by text or cloned text missing selection cue

appears to be related to the following problem:

in the file Layout-TNG-Output.cpp, in the routine Layout::bounds(), there are cases where the routine is called with the function _glyphs.size() = 0. This leads to a failure of the bbox calculation, which leads to this bug. The same problem actually occurs whenever fresh text is typed into a new textbox, except that in this case the results are not visibly wrong. Subsequent calls to this routine increment the size value to 1 or more and all is well. But for the case of a clip this refresh apparently occurs too late.
    In any event, I cannot figure out who is the original caller of this routine and why it is not properly initiallized, so I am going to look for something easier...

Reproduced with Inkscape 0.48+devel r13449 on OS X 10.7.5.

In tests with archived builds on OS X 10.7.5 reproduced with rev >= 12005 (earlier builds show arrows and sel cues, but don't update the visual bbox as expected) - i.e. the same revision seems to trigger the described issue with texts as clip-path as in (similar, but not the same) bug #1322940 (with clones as clip-path).

fix committed to rev 13470

this will affect four cases: text, flowedtext, clone of text, clone of flowedtext

> this will affect four cases: (…), clone of text, (…)

Visual bounding box doesn't update as expected with these steps:

1) create rect
2) create text
3) clone text
4) position clone above rect, clip rect with cloned text
5) edit original text (e.g. insert or add characters)
6) switch to select tool and select the clipped rect

Expected:
Visual bounding box fits visual part of the clipped object tightly.

Actual:
Visual bounding box keeps (cached (?)) size of prior state of clone (or orig text object) and doesn't match the visual part of the clipped rect.

On 2014-07-26 21:43 +0200, ~suv wrote:
>> this will affect four cases: (…), clone of text, (…)
>
> Visual bounding box doesn't update as expected with these steps:

Nevermind - this is an older regression. I'll file a separate report (unless already filed but not found).