TTF import broken in 5.0?

Hello,
Since the version 3.8 Squeak correctly displayed Unicode characters if
glyphs were part of TTF file I imported. But Squeak Squeak 5.0 cannot
do this. For example: 'ąčę' (asciiValues 269, 261, 281) is shown
incorrectly. Bug?
regards,
Vaidotas

Hi Vaidotas,

> On Mar 12, 2016, at 2:26 AM, Vaidotas Didžbalis <[hidden email]> wrote:
>
> Hello,
> Since the version 3.8 Squeak correctly displayed Unicode characters if
> glyphs were part of TTF file I imported. But Squeak Squeak 5.0 cannot
> do this. For example: 'ąčę' (asciiValues 269, 261, 281) is shown
> incorrectly. Bug?
> regards,
> Vaidotas

Can you not narrow this down?  Can you check whether the glyphs are in the font correctly or not to see whether it's an import or a display issue?

_,,,^..^,,,_ (phone)

Hi Vaidotas,

The following fix should allow you to import a TTF font
and display the national characters in Squeak 5.0.

In
        TTFontReader>>#decodeCmapFmtTable:
change
                                i < 256 ifTrue:
                                        [seg last > 0 ifTrue:
to
                                        seg last > 0 ifTrue:
and
                                        cmap at: i + 1 put: code]]].
to
                                        cmap at: i + 1 put: code]].
or revert the method to the version
        'nice 12/27/2009 03:10'
.

In
        TTCFontReader>>#processCharMap:
change
                value = 65535 ifFalse: [ "???"
to
                value > 65535 ifFalse: [ "???"
.

Now import the font and check the characters display as expected.

This is how I would do it:
       

| tmpStyleName tmpFontFileName |

tmpStyleName := #Arial.
tmpFontFileName := '/Library/Fonts/Arial.ttf'.

TTCFont removeStyleName: tmpStyleName.
"Remove the existing font if it exists"

TTCFont newTextStyleFromTTFile: tmpFontFileName.
"Import the new TTF font"

FontChooserTool open.
"Check that font"



Best Regards,
Milan

Hi Milan,
On 17.04.2016, at 14:39, Milan Vavra <[hidden email]> wrote:


As I am probably the one who made that not work,
can you explain that so I understand what went wrong?

Best
        -Tobias

"Hi Tobias,

Strings in Squeak are either ByteString or WideString based on the Unicode codePoints of their characters.
Strings are instances of ByteString if they contain characters with codePoint < 256.
And instances of WideString if they contain characters with codePoint >= 256.
"

| tmpString1 tmpString2 tmpString3 tmpString4 tmpString5 tmpString6 |

$A codePoint = 65.

tmpString1 := String with: $A.
tmpString1 class = ByteString.
tmpString1 size = 1.

tmpString2 := String with: (Character codePoint: 269).
tmpString2 class = WideString.
tmpString2 size = 1.

tmpString3 := tmpString1, tmpString2.
tmpString3 class = WideString.
tmpString3 size = 2.

tmpString4 := tmpString3 copyFrom: 1 to: 1.
tmpString4 class = ByteString.
tmpString4 size = 1.

tmpString5 := tmpString2 squeakToUtf8.
tmpString5 class = ByteString.
tmpString5 size = 2.

tmpString6 := tmpString5 utf8ToSqueak.
tmpString6 class = WideString.
tmpString6 size = 1.



"The national characters mentioned have Unicode codePoint >= 256.

When importing a Unicode font,
the guard
        i < 256 ifTrue: []
in
        TTFontReader>>#decodeCmapFmtTable:
discards all characters
with Unicode codePoint >= 256.

This is undesirable as all those characters mentioned have codePoint >= 256.

The guard
                value > 65535 ifFalse: [ ]
in
        TTCFontReader>>#processCharMap:
discards all characters
with Unicode codePoint > 65535.

This is required as
glyphs size = 65536
.

Most languages have characters with codePoint <= 65535 so this should not be a problem.

Best Regards,
Milan
"