Hello,
I am investigating a text-rendering issue in Aspose.PSD for Python via .NET.
Environment:
- Windows
- Python 3.11
- aspose-psd 26.8.0
- License loaded
- Custom font with family and PostScript name “Germany_doc”
Problem:
After updating an existing TextLayer inside a nested Smart Object,
text containing uppercase Ü or Ö disappears entirely from the rendered
output. ASCII text renders successfully.
Observed results:
- “MÜLLER” and “JÖRG” with Germany_doc: entire text is missing.
- “MULLER” and “JORG” with Germany_doc: visible.
- Text containing ß with Germany_doc: reported visible.
- “Jörg Müller” with another font (“jey”): visible.
Note: this last comparison uses lowercase ö/ü, not uppercase Ö/Ü.
The relevant editing sequence is approximately:
tl.update_text(value)
for portion in tl.text_data.items:
portion.style.font_name = font_name
tl.text_data.update_layer_data()
The application also enables faux_bold for Germany_doc text.
For the affected name layers, the logged properties before and after
editing are:
font_name = "Germany_doc"
font_size = 100.0
tracking = 0
leading = 120.00001
auto_leading = True
faux_bold = True
The input log contains the expected Unicode strings. No exception is
reported during this rendering operation. However, we have not yet
verified the text by reopening the saved leaf PSD.
Font inspection:
- Unicode cmap entries for Ü and Ö are present.
- The same original OTF renders these characters successfully using
Pillow/FreeType. - This does not establish that Aspose selects the same font file or
interprets all font tables identically.
We tried alternative font files and editing approaches, but those tests
were not sufficiently isolated to establish the cause. In particular,
duplicate internal font names may have affected font selection.
Could you please advise:
- Is there a known issue involving uppercase umlauts and faux_bold?
- How can we determine which physical font file Aspose actually uses?
- What diagnostic information should we capture to distinguish text
layout, font rendering, serialization and Smart Object cache issues? - Would you prefer a minimal standalone PSD or a nested Smart Object
example for investigation?
We can provide a neutral sample PSD containing only test text, the
relevant code, expected/actual output, and the font file if redistribution
is permitted.
Thank you.