https://github.com/maximus5/write-read-test

DBCS console difference, gh-739
https://github.com/maximus5/write-read-test

Last synced: 10 months ago
JSON representation

DBCS console difference, gh-739

• Host: GitHub
• URL: https://github.com/maximus5/write-read-test
• Owner: Maximus5
• Created: 2016-06-27T09:54:42.000Z (over 9 years ago)
• Default Branch: master
• Last Pushed: 2016-06-29T08:15:49.000Z (over 9 years ago)
• Last Synced: 2025-01-31T11:50:16.889Z (12 months ago)
• Language: C++
• Size: 21.5 KB
• Stars: 0
• Watchers: 3
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: readme.md

Awesome Lists containing this project

README

          
# This simple test shows crazyness of DBCS processing in Windows console

## Non-DBCS enabled OS (EU, US, RU, etc).

### Raster font is selected in console

* Console can't store or show any characters out of OEM CP. So all CJK glyphs will be displayed as `?`.

* One (broken) CJK glyph takes one cell, so 78 CJK glyphs take same space as 78 ASCII chars.

![screens/non-dbcs-raster.png](https://github.com/Maximus5/Write-Read-Test/raw/master/screens/non-dbcs-raster.png)

### TTF font is selected in console

* Console is working in Unicode mode and **can** store Unicode codepoints, and may display them, if you have configured FontLink.

* One CJK glyph also takes one cell, so 78 CJK glyphs take same space as 78 ASCII chars.

![screens/non-dbcs-unicode.png](https://github.com/Maximus5/Write-Read-Test/raw/master/screens/non-dbcs-unicode.png)

## DBCS enabled OS (Chinese, Japanese, etc).

The crazyness begins. I tested CP 936.

* Console is able to store large set of characters, because CP 936 is double byte character set, and therefore this CP covers wide set of Unicode codepoint.

* NB, doesn't matter what font is selected in console.

### Raster font is selected in console

* One CJK glyph is **displayed** in place of **two cells**.

* Unicode console API returns each CJK in **one** `CHAR_INFO` and do not use `COMMON_LVB_LEADING_BYTE` and `COMMON_LVB_TRAILING_BYTE` flags.

* If console application tries to write more than (console_width/2) CJK glyphs, the string is automatically **folded to next line**.

![screens/dbcs-raster.png](https://github.com/Maximus5/Write-Read-Test/raw/master/screens/dbcs-raster.png)

### TTF font is selected in console

* Same as above, one CJK glyph is **displayed** in place of **two cells**.

* But Unicode console API returns each CJK as **two** consecutive `CHAR_INFO` first with `COMMON_LVB_LEADING_BYTE` (0x100) and second with `COMMON_LVB_TRAILING_BYTE` (0x200) flags.

* Same as above, if console application tries to write more than (console_width/2) CJK glyphs, the string is automatically **folded to next line**.

![screens/dbcs-unicode.png](https://github.com/Maximus5/Write-Read-Test/raw/master/screens/dbcs-unicode.png)

## Bottom line

How it can be bearable?