cjk texts in the pdf file which MkIV output sometimes can not be copy rightly

newer
How to center a table horizontally?

older
Placing boxes anywhere on the page

Li Yanrui (李延瑞)

10 Sep 2010 10 Sep '10

6:30 a.m.

Hi all, For the pdf file which is generated frome the following example, two Chinese characters can not be copied rightly when I use simsun.ttc font and use *Adobe Reader* to view it. The copy text is displayed as the wrong unicode text such as "􁨀". However these glyphs can be copied rightly if I use AdobeSongStd.otf font or I use other pdf reader such as evince. \definefont[song][file:simsun] \starttext \song 描 $\overline{q}$ 提 \stoptext The attachment for this mail is that pdf file with simsun.ttc font which can be downloaded from http://code.google.com/p/way2ctx/downloads/list This situation is not general. For example, if I change one Chinese character the text can be copied normally even using simsun.ttc font. -- Best regards, Li Yanrui (李延瑞)

Attachments:

a.pdf (application/pdf — 11.3 KB)

Show replies by date

Li Yanrui (李延瑞)

10 Sep 10 Sep

6:32 a.m.

New subject: cjk texts in the pdf file which MkIV output sometimes can not be copy rightly

2010/9/10 Li Yanrui (李延瑞) :

...

Hi all,

For the pdf file which is generated frome the following example, two Chinese characters can not be copied rightly when I use simsun.ttc font and use *Adobe Reader* to view it. The copy text is displayed as the wrong unicode text such as "􁨀". However these glyphs can be copied rightly if I use AdobeSongStd.otf font or I use other pdf reader such as evince.

\definefont[song][file:simsun] \starttext \song 描 $\overline{q}$ 提 \stoptext

The attachment for this mail is that pdf file with simsun.ttc font which can be downloaded from http://code.google.com/p/way2ctx/downloads/list

This situation is not general. For example, if I change one Chinese character the text can be copied normally even using simsun.ttc font.

-- Best regards,

Li Yanrui (李延瑞)

My test enviorment is the beta 2010.09.09 23:45 in linux x86. -- Best regards, Li Yanrui (李延瑞)

Taco Hoekwater

6:44 a.m.

New subject: cjk texts in the pdf file which MkIV output sometimes can not be copy rightly

On 09/10/2010 02:32 PM, Li Yanrui (李延瑞) wrote:

...

2010/9/10 Li Yanrui (李延瑞):

...
Hi all,

For the pdf file which is generated frome the following example, two Chinese characters can not be copied rightly when I use simsun.ttc font and use *Adobe Reader* to view it. The copy text is displayed as the wrong unicode text such as "􁨀".

It looks like the reader is ignoring the ToUnicode entry in the bad case. No idea why, though. Best wishes, Taco

Li Yanrui (李延瑞)

20 Sep 20 Sep

5:20 a.m.

New subject: cjk texts in the pdf file which MkIV output sometimes can not be copy rightly

2010/9/10 Taco Hoekwater :

...

On 09/10/2010 02:32 PM, Li Yanrui (李延瑞) wrote:

...
2010/9/10 Li Yanrui (李延瑞):

...
Hi all,

For the pdf file which is generated frome the following example, two Chinese characters can not be copied rightly when I use simsun.ttc font and use *Adobe Reader* to view it. The copy text is displayed as the wrong unicode text such as "􁨀".

It looks like the reader is ignoring the ToUnicode entry in the bad case. No idea why, though.

My friend sent a mail to Ken Lunde who works for adobe systems incorporated. He replied: [quote] I forwarded your email to the Acrobat team for investigation. When I copied the body text from the first file, and all of the text from the second file, the code points are PUA, specifically Unicode Plane 16. The heading text of the first file appears to be encoded correctly. I am guessing that this is a PDF producer issue. I will keep you posted about what the Acrobat team discovers. [/quote] In the above, the first file is the tex file of the pdf file which is the attachment for my previous mail; the second one is that attachment. -- Best regards, Li Yanrui (李延瑞)

Li Yanrui (李延瑞)

16 Nov 16 Nov

5:25 p.m.

New subject: cjk texts in the pdf file which MkIV output sometimes can not be copy rightly

2010/9/20 Li Yanrui (李延瑞) :

...

2010/9/10 Taco Hoekwater :

...
On 09/10/2010 02:32 PM, Li Yanrui (李延瑞) wrote:

...
2010/9/10 Li Yanrui (李延瑞):

...
Hi all,

For the pdf file which is generated frome the following example, two Chinese characters can not be copied rightly when I use simsun.ttc font and use *Adobe Reader* to view it. The copy text is displayed as the wrong unicode text such as "􁨀".

It looks like the reader is ignoring the ToUnicode entry in the bad case. No idea why, though.

My friend sent a mail to Ken Lunde who works for adobe systems incorporated. He replied:

[quote] I forwarded your email to the Acrobat team for investigation. When I copied the body text from the first file, and all of the text from the second file, the code points are PUA, specifically Unicode Plane 16. The heading text of the first file appears to be encoded correctly. I am guessing that this is a PDF producer issue. I will keep you posted about what the Acrobat team discovers. [/quote]

In the above, the first file is the tex file of the pdf file which is the attachment for my previous mail; the second one is that attachment.

Recently Ken Lunde replied again:

...

Gu Hua forwarded your email to the Acrobat, which investigated this issue.

The evidence points toward a malformed or poorly-made ToUnicode table. It would be appropriate for Adobe Reader to ignore such a ToUnicode table.

It would be useful to know how the ToUnicode table is made by this PDF producer. I know that when making a static ToUnicode mapping resource, which uses CMap resource syntax, it is very easy to make an invalid one, specifically because CID ranges, when expressed as two-byte values, cannot cross first-byte boundaries. Also, UTF-32 (and UTF-8) cannot be used directly, and such values must be converted to UTF-16. One or both of these issues, if not handles correctly, can result in an invalid ToUnicode table.

In fact the text with fonts such as simsun.ttc can not be copied rightly but the text with fonts such as AdobeSongStd-Light.otf can be copied rightly. It seems like that this problem may be related to the font loading script of mkiv. -- Best regards, Li Yanrui (李延瑞)

Taco Hoekwater

17 Nov 17 Nov

2:32 a.m.

New subject: cjk texts in the pdf file which MkIV output sometimes can not be copy rightly

On 11/17/2010 01:25 AM, Li Yanrui (李延瑞) wrote:

...

...
In fact the text with fonts such as simsun.ttc can not be copied rightly but the text with fonts such as AdobeSongStd-Light.otf can be copied rightly. It seems like that this problem may be related to the font loading script of mkiv.

That is useful to know. So we should create a pdf file without compression (\nopdfcompression) that contains a small bit of text in both AdobeSongStd and simsun. Can you create an example for that? Then I can compare the two generated CMAPs. Best wishes, Taci

Hans Hagen

2:49 a.m.

New subject: cjk texts in the pdf file which MkIV output sometimes can not be copy rightly

On 17-11-2010 10:32, Taco Hoekwater wrote:

...

On 11/17/2010 01:25 AM, Li Yanrui (李延瑞) wrote:

...
...
In fact the text with fonts such as simsun.ttc can not be copied rightly but the text with fonts such as AdobeSongStd-Light.otf can be copied rightly. It seems like that this problem may be related to the font loading script of mkiv.

That is useful to know. So we should create a pdf file without compression (\nopdfcompression) that contains a small bit of text in both AdobeSongStd and simsun. Can you create an example for that? Then I can compare the two generated CMAPs.

ha, I was keying in the same ... indeed we need an example Hans ----------------------------------------------------------------- Hans Hagen | PRAGMA ADE Ridderstraat 27 | 8061 GH Hasselt | The Netherlands tel: 038 477 53 69 | voip: 087 875 68 74 | www.pragma-ade.com | www.pragma-pod.nl -----------------------------------------------------------------

5000

Age (days ago)

5068

Last active (days ago)

List overview

Download

6 comments

3 participants

participants (3)

Hans Hagen
Li Yanrui (李延瑞)
Taco Hoekwater