SIGN IN SIGN UP

fix: strip soft hyphen when joining merged text elements (#3232)

* fix: strip soft hyphen when joining merged text elements

When predict_merges joins two text elements, _merge_elements always
inserts a space: new_item.text += f' {merged_elem.text}'.

For lines ending with U+00AD (SOFT HYPHEN) this produces 'abge­ rackerten'
instead of 'abgerackerten'. A soft hyphen marks an optional line-break
point in typeset text — the two fragments are halves of one word and must
be joined without a space, with the soft hyphen itself removed.

Check for trailing U+00AD before joining and use direct concatenation
(after stripping the soft hyphen) in that case.

Signed-off-by: stone <frank.schruefer@t-online.de>

* style: apply ruff formatting

Signed-off-by: stone <frank.schruefer@t-online.de>

---------

Signed-off-by: stone <frank.schruefer@t-online.de>
F
Frank-Schruefer committed
827489275e16892ec75f767637de2a9451448441
Parent: 043ed2d
Committed by GitHub <noreply@github.com> on 4/17/2026, 5:15:41 AM