wrong character in koreader

tartanpion · 08-21-2020, 05:33 AM

Code:

<p class="center">*</p>

is sometimes rendered correctly like :
*
and sometimes like
?*
whatever the font
css :

Code:

.center {
  text-align: center;
  text-indent: 0;
}

epub is correct with foliate or calibre viewer and epubcheck ok.

NiLuJe · 08-21-2020, 08:55 AM

Check the debug views when selecting the offending line in KOReader (> View HTML).

tartanpion · 08-21-2020, 09:31 AM

Quote:

Originally Posted by NiLuJe

Check the debug views when selecting the offending line in KOReader (> View HTML).

The good formatting have

Code:

<p class="center">*</p>

The wrong have

Code:

<p class="center">□*</p>

Frenzie · 08-21-2020, 09:38 AM

It's there in the source where it happens.

Firefox doesn't display it, but it shows up in the debugger. WebKit/Blink don't show it in the debugger.

If you throw it in a validator, you get:

Code:

Warning: Saw a control character: U+0099.

After looking it up I'm still not really sure what this Single Graphic Character Introducer (hence SGCI in my screenshot) was ever supposed to do, except that it's some super ancient failed thing.
http://unicode.org/mail-arch/unicode...-m10/0050.html
https://www.unicode.org/L2/L2011/112...ol-aliases.txt

Quote:

0080 PADDING CHARACTER (PAD)
0081 HIGH OCTET PRESET (HOP)
0099 SINGLE GRAPHIC CHARACTER INTRODUCER (SGC)

Those 3 were proposed (on spec) in early drafts of 10646, for what became
a failed architectural direction for 10646. They would be completely forgotten
now except for the persistent (and pernicious) RFC that lists them without
indicating their failed status. Nobody has ever implemented them, so they
are nothing more than character encoding curiosities.

So it seems the correct thing to do would be to ignore these for display, but I remain confused with regard to the presence of these obsolete things from the early to mid '90s.

Frenzie · 08-21-2020, 09:40 AM

It's there in the source where it happens.

Firefox doesn't display it, but it shows up in the debugger. WebKit/Blink don't show it in the debugger (see here).

If you throw it in a validator, you get:

Code:

Warning: Saw a control character: U+0099.

After looking it up I'm still not really sure what this Single Graphic Character Introducer (hence SGCI in my screenshot) was ever supposed to do, except that it's some super ancient failed thing.
http://unicode.org/mail-arch/unicode...-m10/0050.html
https://www.unicode.org/L2/L2011/112...ol-aliases.txt

Quote:

0080 PADDING CHARACTER (PAD)
0081 HIGH OCTET PRESET (HOP)
0099 SINGLE GRAPHIC CHARACTER INTRODUCER (SGC)

Those 3 were proposed (on spec) in early drafts of 10646, for what became
a failed architectural direction for 10646. They would be completely forgotten
now except for the persistent (and pernicious) RFC that lists them without
indicating their failed status. Nobody has ever implemented them, so they
are nothing more than character encoding curiosities.

So it seems the correct thing to do would be to ignore these for display, but I remain confused with regard to the presence of these obsolete things from the early to mid '90s.

tartanpion · 08-21-2020, 10:18 AM

I have made a search and replace with calibre editor. But visually nothing distinguishes the good * from the bad character □*.

NiLuJe · 08-21-2020, 05:42 PM

That's (probably) because Calibre uses Qt5's Webengine, i.e., Blink.
Same deal w/ Sigil.

You'll have to go old-school.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
KOReader en français / KOReader in french	cramoisi	Software	33	04-04-2020 12:14 PM
Name of a character help	skb	Writers' Corner	33	08-21-2018 05:25 PM
The Robbery: A Short Story that goes wrong for all the wrong reasons	brinlingfm	Self-Promotions by Authors and Publishers	0	03-22-2011 08:20 AM
Dates wrong on scheduled news: what am I doing wrong?	Rod Laird	Calibre	5	11-05-2010 06:06 PM
Unutterably Silly Name this character!	AprilHare	Lounge	7	08-23-2009 05:01 PM

08-21-2020, 08:55 AM	#2
NiLuJe BLAM! Posts: 13,506 Karma: 26047190 Join Date: Jun 2010 Location: Paris, France Device: Kindle 2i, 3g, 4, 5w, PW, PW2, PW5; Kobo H2O, Forma, Elipsa, Sage, C2E	Check the debug views when selecting the offending line in KOReader (> View HTML).

08-21-2020, 10:18 AM	#6
tartanpion Enthusiast Posts: 30 Karma: 10 Join Date: Nov 2011 Location: france Device: kobo libra 2	I have made a search and replace with calibre editor. But visually nothing distinguishes the good * from the bad character □*.

08-21-2020, 05:42 PM	#7
NiLuJe BLAM! Posts: 13,506 Karma: 26047190 Join Date: Jun 2010 Location: Paris, France Device: Kindle 2i, 3g, 4, 5w, PW, PW2, PW5; Kobo H2O, Forma, Elipsa, Sage, C2E	That's (probably) because Calibre uses Qt5's Webengine, i.e., Blink. Same deal w/ Sigil. You'll have to go old-school.