LibreDraw shows missing fonts from pdf when the encoding is Identity-H

[mchsm19]

Getting same issue as per reported on
https://ask.libreoffice.org/t/libreoffice-does-not-recognize-identity-h-fonts/24885

Steps to reproduce behavior:

1. Open edge/chrome/firefox browser
2. Navigate to https://documentation.solarwinds.com/en/success_center/orionplatform/content/core-deploy-orion-in-multi-tenant-way.htm
3. Print and select printer “Save as PDF”
4. Open the saved pdf above in libredraw
5. Result: all text with character f will be missing (e.g. Con igure, Speci y, Plat orm, di erent)

(if we open it with pdf reader, the f character will be shown properly)

From pdf properties, we can see the document encoding is Identity-H.
Any advice if LO can resolve this issue?

Thanks.

[mikekaganski]

LibreOffice does not use fonts embedded in PDF. It also does not recode characters - it assumes Unicode character encoding. In an Identity-H encoding, ligatures (like ff, fi, ti, etc.) have own custom codes still, so LibreOffice will be unable to map them to something proper.

Remember: LibreOffice is not a proper PDF viewer / editor.

[mchsm19]

Is there plan for future improvement on it?

LO is not a pdf viewer / editor, but import function is one of the features there.

The import feature will look like broken/bug when the pdf embeds those kinds of fonts

[mikekaganski]

No plans. But any contribution is welcome: doers decide.

Note though, that using an embedded font with a partial subset would be ultimately impossible anyway: LibreOffice needs normal font…
Think about it. A custom encoding means to PDF reader: take the “image” of the character from the embedded font’s character X (where X is some random position, not some sensible/known encoding). The reader has no idea what that character is (i.e., it can’t tell you e.g. which Unicode is that; try to copy the text from a PDF reader that shows it to you normally, and paste into any text editor, to see that) - but it can show it to you, using that font, as it is intended.

OTOH, LibreOffice needs to map all characters in any imported file to Unicode. And if there is no information in the file how to do that, it will be unable to. End of story; and “will look like broken/bug” is just a misconception - because ultimately, PDF is not designed to be editable - it is a “hard copy”, a snapshot of some state of a document at some fixed point in time, and editing / importing it is wrong conceptually, even though common.