Text Format Problem When Opening PDF in Beta Versions

[FullGrim]

Greetings. I noticed a new bug in the current beta versions (I mostly use Photo, but the same problem seems to exist in Designer and Publisher). When I open a simple PDF file (written in Word) in version 2.5.7, it opens without any problems as follows:

When I open the same PDF file with the same settings in the current beta versions, all the lowercase letters are converted to uppercase and the text format is completely corrupted, as follows:

I have tried repairing beta apps, resetting, uninstalling and then deleting all Affinity folders and reinstalling, etc. but I can't fix the problem.

By the way, as a suggestion; It would be great if an option could be added to the PDF opening window to open the PDF as a non-editable flat image (without the risk of corruption the text format or page design). For example; It would be very useful if we want to open all the pages of a complex 300-page PDF file in Photo and export them separately in png format in Export Persona without making any changes to them.

My system: CPU: i3-12100F, RAM: 32 GB DDR 4 3200, GPU: RTX 4060, SSD: Gen3 M2, OS: Win 10 Pro (22H2), all apps MSIX setup

Edited January 12 by FullGrim
typo

[kenmcd]

Can you attach the PDF?

[MikeTO]

Could you share one page of the PDF here for testing? I'm unable to duplicate this with my own PDF files.

Does it happen with the "Favour Editable Text Over Fidelity" checkbox on and off?

[FullGrim]

I have attached the sample PDF file.

I am trying all the option combinations in the PDF opening window but there is no change. By the way, although the file uses only Times New Roman font, it gives a warning that Arial fonts are missing (in version 2.5.7, it also gives a warning that Arial is missing but it opens properly).

sample3.pdf

[MikeTO]

FWIW it opens correctly on macOS.

• The PDF defines four fonts: Arial, Arial Bold, Times New Roman, and Times New Roman Bold. I don't understand why Affinity thinks these fonts are missing.
• It would be nice if 2.6.0 could replace un-styled versions of Arial and Times with Arial Regular and Times Regular like 2.5.7 does without prompting.
• I don't know why 2.6.0 wants to replace Arial Bold with Helvetica Bold, and Times Bold with Helvetica Bold since it is replacing the un-styled versions with Arial and Times, respectively.
• Although these four fonts are defined in the PDF, all the text is formatted as Times (no style) so only the 3rd font in the 2.6.0 dialog matters.
• The replacement family choices for the two Bold styles in 2.6.0 are sticky when opening the PDF multiple times, just as they were in 2.5.7, but the choices for the two no style fonts aren't sticky. I don't really care either way, I just found it surprising.

2.5.7 (left), 2.6.0 3027 (right)

[anto]

13 hours ago, FullGrim said:

all apps MSIX setup

On my windows 10 where i have all exe programs file opens correctly. Maybe problem in msix?

[Hangman]

Just now, anto said:

On my windows 10 where i have all exe programs file opens correctly.

Same here...

[MikeTO]

Serif, this is unrelated to the issue being discussed in this thread, but the test document above is another one that opens as Unknown (de-DE) for me. This has become very common with 2.5.x, I'm seeing it every couple of days with somebody's test document. If the app knows it's de-DE and the German dictionary is installed, why is it Unknown?

The other documents I've seen this with include fr-FR and en-US.

[Obscured]

.

[Hangman]

34 minutes ago, MikeTO said:

Serif, this is unrelated to the issue being discussed in this thread, but the test document above is another one that opens as Unknown (de-DE) for me. This has become very common with 2.5.x, I'm seeing it every couple of days with somebody's test document. If the app knows it's de-DE and the German dictionary is installed, why is it Unknown?

For me, it opens with English (United Kingdom) on both Mac and Windows...

[FullGrim]

Thanks. I uninstalled the MSIX installations of the beta apps and installed the MSI/EXE versions, but there was no change. I also ran diagnostics like sfc /scannow in Windows, still no change. Interesting, I think this problem is not general, it is just specific to my system. I hope I will not have problems when version 2.5.7 is updated to 2.6, because I have to work with PDFs all the time.

2 hours ago, MikeTO said:

Serif, this is unrelated to the issue being discussed in this thread, but the test document above is another one that opens as Unknown (de-DE) for me. This has become very common with 2.5.x, I'm seeing it every couple of days with somebody's test document. If the app knows it's de-DE and the German dictionary is installed, why is it Unknown?

The other documents I've seen this with include fr-FR and en-US.

English and Turkish languages are installed on my system, so I think the sample PDF file I uploaded may be identify as an unknown language on your system.

[MikeTO]

2 hours ago, Hangman said:

For me, it opens with English (United Kingdom) on both Mac and Windows...

Hmm, it is for me now, too. But it was like I showed when I took that screenshot. Odd.

[kenmcd]

5 hours ago, MikeTO said:

I don't know why 2.6.0 wants to replace Arial Bold with Helvetica Bold, and Times Bold with Helvetica Bold since it is replacing the un-styled versions with Arial and Times, respectively.

Affinity still cannot handle the comma in the font name when identifying fonts.
Such as: Arial,Bold or Arial,Italic or Times New Roman,Bold
Awhile back I went to find where these font names with the commas came from.
According the Microsoft .NET API the comma tells the text rasterizer to use the Bold font if available, or to apply the fake Bold if the actual font is not available.
The sample document above was created with Microsoft Word 2007.
When a user prints the doc and they do not have the Bold font, it still looks bold.
So it is a way of MS handling the fake bold and fake italics.

The Affinity font matching code should know that:
- Arial,Bold = Arial-Bold 
- Arial,Italic = Arial-Italic
- etc.

Arial and Times New Roman are kinda funky with the names embedded.
But many other fonts the FontName,Bold = FontName-Bold exactly in the PS Name.
So there is no excuse to not match them correctly.

This has come-up multiple times before in other discussions about matching fonts.