z@*.tmp coming up as missing font in pdf

[dianne32]

Hi, I have tried searching but have not found any questions regarding this issue. When I opened up a pdf document in affinity publisher Z@*.tmp comes up as a missing font (several of these come up such as z@rac78.tmp). I have a font to replace it with but words come up as gibberish in the document. I tried deleting my temporary files, restarting the computer but it is still an issue. When I open the same file in adobe acrobat, the pdf file looks fine. It is driving me crazy, LOL. Any help or suggestions are appreciated. Thanks so much. 

[v_kyr]

5 hours ago, dianne32 said:

When I open the same file in adobe acrobat, the pdf file looks fine.

Then check in Acrobat which fonts have been finally used there for that PDF file.

See also related ...

• "Z@R828.tmp" font instead of Arial
• Z@R####.tmp font is displaying when opening a PDF from Outlook
• ... etc.

[kenmcd]

12 hours ago, dianne32 said:

When I open the same file in adobe acrobat, the pdf file looks fine.

Most likely it is simply displaying the shapes that are there - which look like text, but actually there is no embedded text or font.

Or it could actually be text using an encoding that APub does not support. If that is the case it may be possible to recover the text.

Have to see the PDF.

[dianne32]

15 hours ago, v_kyr said:

Then check in Acrobat which fonts have been finally used there for that PDF file.

See also related ...

• "Z@R828.tmp" font instead of Arial
• Z@R####.tmp font is displaying when opening a PDF from Outlook
• ... etc.

Thank you. I will check that. I have read those other links you provided but still haven't been able to fix the issue in affinity. That's where I got the info to delete the temp files. 

 

[dianne32]

8 hours ago, kenmcd said:

Most likely it is simply displaying the shapes that are there - which look like text, but actually there is no embedded text or font.

Or it could actually be text using an encoding that APub does not support. If that is the case it may be possible to recover the text.

Have to see the PDF.

Thanks. I will google recovering the text. 

[kenmcd]

2 hours ago, dianne32 said:

Thanks. I will google recovering the text. 

Gotta know what is actually in there before you can try to recover it. Doubt G will help.

[walt.farrell]

2 hours ago, dianne32 said:

Thanks. I will google recovering the text. 

Can you share the PDF?

What application created it?

[dianne32]

On 7/14/2023 at 11:10 AM, walt.farrell said:

Can you share the PDF?

What application created it?

It is a puzzle file made by someone on fiverr. I am attaching the demo he gave me.  It has the same temp files coming up but final result isn't as gibberish. But it is the same problem. I will get in touch with the creator and see what he used to make it. Thanks for any help.

DEMO Easy-Medium-Hard-Extra Hard Sudoku Puzzles with Solutions.pdf

[walt.farrell]

1 hour ago, dianne32 said:

Thanks for any help.

The reason you're seeing those is that whatever application he used created the PDF that way. Acrobat Reader, for example, shows some embedded Roboto fonts, but also these at the end of the list of fonts for that demo file:

As the Affinity apps cannot use embedded fonts, that approach to creating PDFs won't work with any Affinity app.

[kenmcd]

1 hour ago, dianne32 said:

I am attaching the demo he gave me. 

The only pages that have any actual text are:
- page 2 - "This book belongs to"
- page 3 - the table of contents

Everything else is just shapes.
So all the "text" on Page 1 (page 4) is shapes - there is no actual text.
Same with all the other pages to the end. Just shapes. No text.

And the encoding on the actual text is WinAnsi - which is a pre-Unicode encoding which Affinity applications do not support.
A work-around for this type of actual text with old encoding is to open the file in an application which does understand this old encoding - and these applications will often automatically convert/update this old encoding to Unicode - and then paste that updated Unicode text into Affinity.

The only thing you can do with the shapes "text" is use OCR to convert it to actual text.

[kenmcd]

1 hour ago, dianne32 said:

I will get in touch with the creator and see what he used to make it.

The PDF says it was created with: Adobe Illustrator 27.2 (Windows)

[walt.farrell]

40 minutes ago, kenmcd said:

The PDF says it was created with: Adobe Illustrator 27.2 (Windows)

I wonder if those odd fonts were instantiated/created to handle text that was made from a Variable font?

[kenmcd]

4 hours ago, walt.farrell said:

I wonder if those odd fonts were instantiated/created to handle text that was made from a Variable font?

I don't think so.
Adobe has published guidelines on how they think variable font instances should be named - the PS names used in embedding the fonts.

Quote

Adobe Technical Note #5902
Generating PostScript Names for Fonts Using OpenType Font Variations
Version 1.0, September 14 2016

If it is a named instance, then that name is what is going to be used to embed the font - just like static fonts. In the part about "arbitrary instances" they have guidelines on how to construct PS names to use for embedding custom instances.
And those PS name guidelines look nothing like these .tmp font names.

So I have no idea where these weird .tmp font names are coming from.
Or even why they are there.
There does not appear to be any fonts embedded with those names.

I wonder if the original source was from another application and that was then edited in Illustrator and then output to PDF again. Dunno.

Note: I have seen them embed variable fonts in a manner which does not seem to follow their own guidelines, but still nothing like this.

Another thought, if Adopey had actually done this - there would actually be a font embedded there. They normally do not have any problems embedding variable fonts.

[dianne32]

Thank you so much for the help. I have to admit that this is over my head, lol. I am going to ask the person who created it to look at this and see if we can come up with a solution. Thank you so much.

[kenmcd]

Hmmm... on looking a little closer...
There three .tmp fonts do actually have characters embedded.
Looks like they are used in the headings on some of the tables.
And they are encoded as Unicode.
And the characters look like they could actually be from Roboto-Bold.

The five styles of Roboto which are embedded are all WinAnsi encoding.
Why would some fonts be embedded as Unicode and some as WinAnsi?
Makes no sense.

Gonna see if I can replace all the fonts...

[dianne32]

I got in touch with the person who made the file. I showed him this thread. He ended up changing the font to Roboto and that worked. Yeah! Thanks so much for all the suggestions!

[thomaso]

On 7/15/2023 at 7:03 PM, kenmcd said:

The only pages that have any actual text are:
- page 2 - "This book belongs to"
- page 3 - the table of contents

Everything else is just shapes.
So all the "text" on Page 1 (page 4) is shapes - there is no actual text.
Same with all the other pages to the end. Just shapes. No text.

There seem to be some more text layers across the entire document.
Unfortunately many of the text frames are cropped by a nested vector mask. – Does someone have an idea what workflow in the initial layout would cause these masks in the PDF / for Affinity?

---

[walt.farrell]

29 minutes ago, thomaso said:

Does someone have an idea what workflow in the initial layout would cause these masks in the PDF / for Affinity?

From earlier posts, it was created in Illustrator, not one of the Affinity applications.

[thomaso]

27 minutes ago, walt.farrell said:

From earlier posts, it was created in Illustrator, not one of the Affinity applications.

That's what my first screenshot says. too. – But what workflow (in AI, or on PDF export or conversion) might cause the masked text frames in this PDF? It is rather quite unlikely that they got masked by the initial designer / in the Illustrator layout document, right? But I can't think of any export option that causes such masks for text either.

A possibly related question would be: How come that on the pages with tables (?) the numbers got curved but the texts above/below did not ('instead' they got masked)?

Even in case the initial designer imported those items from a third application into Illustrator, I still wonder what step exactly would cause this, and with what goal (to prevent / achieve what)?

EDIT: obviously the masks exist in the PDF as white filled paths:

Edited July 17, 2023 by thomaso

[v_kyr]

50 minutes ago, thomaso said:

But what workflow (in AI, or on PDF export or conversion) might cause the masked text frames in this PDF? It is rather quite unlikely that they got masked by the initial designer / in the Illustrator layout document, right? But I can't think of any export option that causes such masks for text either.

Read through this here to get a vague idea ...

• Z@xxx.tmp files in user temp folder causes gibberish font in printout

... the same will happen for PDF generation then (instead of print outs) and by reusing such generated PDFs then in other PDF format related apps.

Further a bunch of Sudoku Generators (Python, JS, PHP, ... based etc.) on the net do also offer to generate PDFs directly as output, they make therefor use of some third party PDF generation libs, where some of those in turn have some own bugs/problems with certain font usages here too when generating PDF files (...some of them especially with certain Arial and Roboto fonts).

[thomaso]

4 minutes ago, v_kyr said:

Read through this here to get a vague idea ...

It seems to explain the culprit / the font issue – but does it mention "mask"? Whereas, yes, an involved Sudoku Generator with a buggy PDF generation lib could be reason. Thanks.

But I am still curious to understand for what purpose such a command gets written (~ "create a mask for text frames in similar size but slightly larger"), either in a Sudoku or a PDF generator lib. I just can't imagine any use for this, especially because these masks don't really mask anything visually. (the do in my screenshots only because of my font replacement)

[kenmcd]

5 hours ago, thomaso said:

There seem to be some more text layers across the entire document.

As mentioned above...

22 hours ago, kenmcd said:

Hmmm... on looking a little closer...
There three .tmp fonts do actually have characters embedded.
Looks like they are used in the headings on some of the tables.
And they are encoded as Unicode.
And the characters look like they could actually be from Roboto-Bold.

The five styles of Roboto which are embedded are all WinAnsi encoding.
Why would some fonts be embedded as Unicode and some as WinAnsi?
Makes no sense.