Line breaking in the simplified chinese pages

[jrshi]

Here is the line-breaking bug when viewing a Chinese page encoded in GB2312 or GB18030.

* Version of software in which the issue occurs

-- iSilo 4.01-4.05 for Palm
-- ISiloX for win 4.01-4.05

* Exact steps to reproduce the issue

1. Convert the following page: http://www.guoxue.com/shibu/24shi/shiji/sj_001.htm
2. View the page in iSilo for palm. -- See the attached screenshot for
the result.


* Operating system version
-- Palm OS5.2.1

* PDA model

-- Sony Clie TH55

* Special hardware configuration

-- None

* Special installed third-party software, especially those which integrate at the system level

-- The Chinese is supported by CJKOS 4.61. The default locale is set to GBK (GB18030).
---------------------------------------------------------------------
The problem is:

Chinese characters are encoded in two-bytes. But in the case shown in the attached screenshot, the iSilo viewer breaks the line at the middle of a chinese character (one byte at the end of the current line, and another byte at the beginning of the next line), rendering the chinese characters following it totally messed up.

It seems the wrong line-break only occurs at the certain places. As shown in the screenshot attached, it happens when the text just switches to a new environment (eg. switch to another font/color) and its first character happens to reside at the end of the line -- see the attached image and you will know what I mean. I have checked the whole document, all the wrong line breaks happen at same particular places.

If this bug can be fixed, I guess the chinese display will be nearly perfect. There had been a lot complains about the problem in various chinese forums -- I guess this had prevent many chinese users from adopting iSilo as their primary palm reader.

[iSilo]

Scanning through the page at the URL you gave, I wasn't able to locate the text shown in your screen shot. Perhaps you could repost the screen shot with the scroll bar enabled and tell us which paragraph the text is in.

[jrshi]

Sorry. The page shown in the screenshot was actually taken from
http://www.guoxue.com/shibu/24shi/shiji/sj_002.htm
-- the messy chinese display demonstated in the screenshot can be found in all sj_*.htm.

I include another screenshot -- now the scrollbar is turn on and the messy display is in the first paragraph. Please note that the exact places where the messy display occurs depend on the display setting (eg. font size, scrollbar on/off, viewer version and conversion setting etc.). For instance, the messy display shown in this new screenshot is different from the original one becuase I turn on the scrollbar and the text positioning is changed, and the original one now displays perfectly normal. For the particular messy display shown in this screenshot, you may find it displayed perfectly normal in your system because the setting maybe different -- but you can always find messy display in some other places (as a rule of thumb, non-ascii character such as AE, Sterling symbol etc. shows up when there is messy display -- the correct display will not include any western symbols).

Thanks for the attention.

Scanning through the page at the URL you gave, I wasn't able to locate the text shown in your screen shot. Perhaps you could repost the screen shot with the scroll bar enabled and tell us which paragraph the text is in.

[iSilo]

Thanks for the additional details. The issue is under investigation.

[iSilo]

Version 4.15, which has just entered the beta1 (http://forum.isilo.com/showthread.php?t=626) test stage, fixes this issue. Specifically, the issue is that with CJKOS, the character at the end of a line may display incorrectly if it begins a new formatting run.