Gustav Andersson created FOP-2657:
-------------------------------------

             Summary: Certain custom fonts are shifted 'up' in the PDF due to 
wrong ascender metric
                 Key: FOP-2657
                 URL: https://issues.apache.org/jira/browse/FOP-2657
             Project: FOP
          Issue Type: Bug
          Components: font/unqualified
    Affects Versions: trunk
            Reporter: Gustav Andersson


Certain fonts are shifted 'up' in the PDFs. I've noticed this in the Google 
Noto Symbol and Naskh Arabic fonts, but not in the standard Google Noto Sans 
font.

If I generate the metric XML file for the three fonts, and copy the value of 
the <ascender> and <descender> element from the Noto Sans font to the metric 
files generated for the Symbol and Naskh Arabic fonts, those two fonts are then 
'fixed' and appear in the correct place again. 

(Not that the Naskh Arabic font is overflowing its line at the bottom, so there 
is still something wrong there, but a lot better.)

To reproduce:
# Download the 
[NotoSans-Regular.ttf|https://www.google.com/get/noto/#sans-lgc], 
[NotoNaskhArabic-Regular.ttf|https://www.google.com/get/noto/#naskh-arab] and 
[NotoSansSymbols-Regular.ttf|https://www.google.com/get/noto/#sans-zsym] font 
files and put in a folder.
# Download the attached [^input.fo], [^FOPconfig.xml], [^NotoSans-Regular.xml], 
[^NotoNaskhArabic-Regular.xml] and [^NotoSansSymbols-Regular.xml] and put them 
in the same folder.
# Run {{fop -c FOPconfig.xml input.fo output.pdf}}
# The output should look like the attached [^output.pdf] file, showing the 
issue.

The three metric XML files were generated using the instructions on the Apache 
FOP Confgiuration page, and I then edited the NotoNaskhArabic-Regular.ttf and 
NotoSansSymbols-Regular.ttf by overwriting the <ascender> and <descender> 
elements with those found in the NotoSans-Regular.xml. Note that I don't know 
anything about the metric files, so I don't know if editing those elements were 
the correct thing to do, or even what they mean!



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to