Affinity Publisher: know which Unicode character is selected

[Pyanepsion]

Hello everyone.

Sometimes the text copied into Affinity Publisher contains Unicode characters that do not exist in the font used. How can I find out what the Unicode number of this character is?

Example: It looks like a [à], but it is actually a combination of a dead letter [grave accent] followed by a lower case a. What is this particular character or these two characters? How to find it?

Thank you for your explanations.

🔴 Note. Once again, the French translation is more than just a mess. This time, the translators of the French team didn’t even check if the translation appeared in full!

‘Unsupported characters used’ should be translated as ‘Utilisation de caractères non supportés’. Here, what is written means absolutely nothing.

[garrettm30]

29 minutes ago, Pyanepsion said:

How can I find out what the Unicode number of this character is?

Select the text, then use the menu item Text->Toggle Unicode. I think that will give you what you are asking.

[Pyanepsion]

Thanks, @garrettm30.

Well, this feature seems to be difficult in this case. See the attached file.

🔴 By the way. Another translation error by the French translation team. They translated "Toogle Unicode" as 'Activer/Désactiver Unicode" (Enable/Disable Unicode) which has absolutely nothing to do with! And is in fact impossible.

a-grave.afpub

[garrettm30]

7 minutes ago, Pyanepsion said:

See the attached file.

In that file, the grave accent was already decomposed from the letter. Was that intended?

In any case, the toggle just renders the code as text, so the extreme font size of the text made the code points hard to see due to overflow.

[garrettm30]

3 minutes ago, garrettm30 said:

In that file, the grave accent was already decomposed from the letter. Was that intended?

Actually, I discovered that the accent is a missing character in my version of Garamond. When I changed fonts, I found that the accent was composed with the a properly. Then it gives the value of U+0061U+0300.

Note that it was a little hard to make the selection to include the composed accent. The first time I tried, it appears to have selected just the letter, giving me this result (notice the accent is squished into the 1 in the code):

I think what could be improved here is that text selection of composed characters should select both parts of composed character.

 

[Pyanepsion]

My enlarged example is of course taken from the text shown as a screenshot in the first post. Here the author has used several operating systems, hence the accented letters sometimes consisting of two letters, sometimes the correct character, and sometimes a circumflex dead letter. There are also characters that seem to have a particular functionality, for example perhaps what the Binder would be without a hunt, a U+200D (not sure, as it seems impossible to select it in the text with the mouse).

[Wosven]

I've already seen such problems with .doc documents.  It's not easy to hunt all those characters. Sometimes, they can appears as another language, and this can help to find them.

Here, it would be helpfull to have the code once between tags, so it can be selected only as one character. But this way, you can also search only "U+"

Courage !

[kenmcd]

Those two characters (U+0061U and U+0300) should not be there as two characters,
it should be one character (U-00E0, Latin Small Letter A With Grave).
Find and Replace works. I tested it.

Old fonts often had these separate diacritics such as U+0300 which users could enter manually.
This was a way for old fonts with only 256 characters to support many other accented characters.
But the most obvious problem is the one-size-fits-all placement can be way off.
In modern fonts these characters are pre-composed by the font designer, and the diacritics are placed exactly where they want them.
Some fonts still have these old characters available, but many do not.

Whoever created this text manually entered those old grave characters.
So you should be able to search for the two characters, and replace them with the current single character.
Should work for a, e, i, etc. and any other combinations.

[Pyanepsion]

5 hours ago, LibreTraining said:

So you should be able to search for the two characters, and replace them with the current single character.

Not really. To clean up the text I first had to search for unrecognized characters using the Font Manager menu¹, and for each new group of characters I found I had to successively cut and replace everything, but for some groups of characters like here I had to select a whole block of letters before and after the sequence at each occurrence so I could delete it and rewrite it entirely by hand. This is obviously time consuming and tedious.

🔴¹ By the way, again and again translation errors by the French translation team. ‘Font Manager’ should be translated into French as ‘Gestionnaire de fontes’ and not ‘Gestionnaire de polices’. A ‘Police’ is a family of fonts (like Arial). The window displays the fonts (like Arial regular, Arial bold, etc.), not the families.

@Wosven It was indeed a .doc file. Being a French university professor sometimes in Asia, and sometimes in France, it was written on different Linux, Mac and Windows systems with different textors. He also in other places used Asian characters that looked like Latin characters (e.g. some kind of brackets [and] that were not).

[Wosven]

We have also those problems when the writers travel a lot or simply with people living in other countries but writing in French on computers, I suppose, set in another language.

[kenmcd]

6 hours ago, Pyanepsion said:

Not really. To clean up the text I first had to search for unrecognized characters using the Font Manager menu¹, and for each new group of characters I found I had to successively cut and replace everything, but for some groups of characters like here I had to select a whole block of letters before and after the sequence at each occurrence so I could delete it and rewrite it entirely by hand. This is obviously time consuming and tedious.

Even pre-Unicode character errors typically have a pattern which can be find/replaced. And typically most of those errors are repeated so find/replace can fix most of them. 

I you can provide a sample doc with these problem areas I would be happy to help.

[Alfred]

8 hours ago, Pyanepsion said:

‘Font Manager’ should be translated into French as ‘Gestionnaire de fontes’ and not ‘Gestionnaire de polices’. A ‘Police’ is a family of fonts (like Arial).

Even among native English speakers there is confusion as to whether an individual font or an entire font family should be called a ‘typeface’.

[Petar Petrenko]

On 9/17/2021 at 4:26 PM, Alfred said:

Even among native English speakers there is confusion as to whether an individual font or an entire font family should be called a ‘typeface’.

I've read somewhere that "Typeface" should be used for the family name (Arial, Courier...) and "Font" for its family members (Bold, Italic...).