KOI8-R in Win95 Netscape

Netscape v3.0-3.01 tuning

Netscape v3.0-3.01 have almost working KOI8-R support (mostly as result of my discussion with Netscape Team, but they forget to implement KOI8-R emulation via CP1251 for end users). To activate KOI8-R support you need KOI8-R fonts (you can get some old-style Win3.1 KOI8-R fonts or use my KOI8-R GDI.EXE fix (preferred)).

Choose Cyrillic(KOI8-R) encoding into Options|General Preferences|Fonts|For the Encoding and proper script for each of two fonts:

For my KOI8-R GDI.EXE fix:
Russian (KOI8-R) (or Русский (KOI8-R) for Win95 Russian Edition)
For old-style Win3.1 KOI8-R fonts:
Western (or Западная Европа for Win95 Russian Edition)
Do not set Cyrillic(KOI8-R) in Options|Document Encoding as permanent default encoding scheme (don't press Set Default). It violates HTML and HTTP standards which says that default document encoding (without any special instructions) must be ISO8859-1. Temporary setting of it is possible as workaround for broken Russian pages or for sending E-mail/News messages.

Bug list:

The correct character translation way for Windows will be decode from KOI8-R to CP1251 for all input and decode from CP1251 to KOI8-R for all output with optional UNICODE fonts usage. Netscape not follow this way and provides raw KOI8-R mode, so users forced to setup KOI8-R fonts and keyboard. But even in raw KOI8-R mode it have following bugs.

According to Internationalization of the Hypertext Markup Language (RFC 2070) charset= from HTTP header must overwrite charset= from <META ...> tag. Netscape does just opposite thing, i.e. <META ...> overwrites HTTP header charset. Load this test page with different HTTP header and META charsets to see this bug in action (check View|Document Info for Charset value, there must be KOI8-R, not windows-1251).
When Cyrillic(KOI8-R) is set as Document Encoding (or it is default encoding by Set Default), most of text-oriented Java applets works incorrectly, i.e. it looks like <PARAM> tags not passed to applet at all. NullPointerException becomes most frequent error. It heals by setting Document Encoding to Western(Latin1), but it is only workaround of real bug. BTW, MS Internet Explorer works right in this case.
This version can't correctly display some special HTML characters such as:
© Copyright character
- <LI> mark character
they taken from ISO8859-1 instead of KOI8-R.
It uses wrong font (CP1251 instead of KOI8-R) for ALT=... image attributes
It uses wrong font (CP1251 instead of KOI8-R) for document <TITLE>s.

Enter your name, E-mail address and send the bug report above to Netscape team, the more bug reports they receive - the higher chances there are that they fix it.

Your Full Name:
E-mail Address:

You can additionly edit this text if you want:

Subject:

Bug report:
MS Windows Netscape 3.01 bugs with Cyrillic(KOI8-R) encoding: The correct character translation way for Windows will be decode from KOI8-R to CP1251 for all input and decode from CP1251 to KOI8-R for all output with optional UNICODE fonts usage. Netscape not follow this way and provides raw KOI8-R mode, so users forced to setup KOI8-R fonts and keyboard. But even in raw KOI8-R mode it have following bugs. 1) According to Internationalization of the Hypertext Markup Language (RFC 2070) charset= from HTTP header must overwrite charset= from <META ...> tag. Netscape does just opposite thing, i.e. <META> overwrites HTTP header charset. Load this test page at http://www.nagual.pp.ru/~ache/a.html8 with different HTTP header and META charsets to see this bug in action (check View|Document Info for Charset value, there must be KOI8-R, not windows-1251). 2) When Cyrillic(KOI8-R) is set as Document Encoding (or it is default encoding by Set Default), most of text-oriented Java applets works incorrectly, i.e. it looks like <PARAM> tags not passed to applet at all. NullPointerException becomes most frequent error. It heals by setting Document Encoding to Western(Latin1), but it is only workaround, please fix this bug. BTW, MS Internet Explorer works right in this case. Look at http://multiserver.kuai.se/Products/NewsLine/NewsLine.html to see this bug in action, among many others from http://www.gamelan.com/pages/Gamelan.fx.text.html 3) Netscape shows wrong special characters for ©, ®,  and list bullets. For those characters iso8859-1 encoding _always_ used instead of KOI8-R encoding (which is different for those characters), please fix this bug. Look at http://www.nagual.pp.ru/~ache/special.html to see this bug in action. 4) Netscape shows ALT="..." image attributes using wrong font (default Windows font, not KOI8-R as specified in document). It makes impossible to use Russian (KOI8-R) text in ALT="...", please fix it. Look at http://www.nagual.pp.ru/~ache/header_r.html to see this bug in action. 5) Netscape shows document <TITLE>s in top of window using wrong font (default Windows font, not KOI8-R as specified in document). It makes documents with Russian <TITLES>s looks ugly. Look at http://www.nagual.pp.ru/~ache/title.html to see this bug in action. For all this tests you need Russian KOI8-R (not Windows default) fonts, of course, according to RFC 1489 specs. Some of public KOI8-R fonts spreaded now not conforms specs because they attempt to workaround your bug(2) by moving special characters from valid KOI8-R places to icorrect (iso8859-1) places.

You can use original Netscape bug report page instead, just Copy this bug report and Paste it there.

Subject:
Bug report:
MS Windows Netscape 3.01 bugs with Cyrillic(KOI8-R) encoding: The correct character translation way for Windows will be decode from KOI8-R to CP1251 for all input and decode from CP1251 to KOI8-R for all output with optional UNICODE fonts usage. Netscape not follow this way and provides raw KOI8-R mode, so users forced to setup KOI8-R fonts and keyboard. But even in raw KOI8-R mode it have following bugs. 1) According to Internationalization of the Hypertext Markup Language (RFC 2070) charset= from HTTP header must overwrite charset= from <META ...> tag. Netscape does just opposite thing, i.e. <META> overwrites HTTP header charset. Load this test page at http://www.nagual.pp.ru/~ache/a.html8 with different HTTP header and META charsets to see this bug in action (check View\|Document Info for Charset value, there must be KOI8-R, not windows-1251). 2) When Cyrillic(KOI8-R) is set as Document Encoding (or it is default encoding by Set Default), most of text-oriented Java applets works incorrectly, i.e. it looks like <PARAM> tags not passed to applet at all. NullPointerException becomes most frequent error. It heals by setting Document Encoding to Western(Latin1), but it is only workaround, please fix this bug. BTW, MS Internet Explorer works right in this case. Look at http://multiserver.kuai.se/Products/NewsLine/NewsLine.html to see this bug in action, among many others from http://www.gamelan.com/pages/Gamelan.fx.text.html 3) Netscape shows wrong special characters for ©, ®, and list bullets. For those characters iso8859-1 encoding _always_ used instead of KOI8-R encoding (which is different for those characters), please fix this bug. Look at http://www.nagual.pp.ru/~ache/special.html to see this bug in action. 4) Netscape shows ALT="..." image attributes using wrong font (default Windows font, not KOI8-R as specified in document). It makes impossible to use Russian (KOI8-R) text in ALT="...", please fix it. Look at http://www.nagual.pp.ru/~ache/header_r.html to see this bug in action. 5) Netscape shows document <TITLE>s in top of window using wrong font (default Windows font, not KOI8-R as specified in document). It makes documents with Russian <TITLES>s looks ugly. Look at http://www.nagual.pp.ru/~ache/title.html to see this bug in action. For all this tests you need Russian KOI8-R (not Windows default) fonts, of course, according to RFC 1489 specs. Some of public KOI8-R fonts spreaded now not conforms specs because they attempt to workaround your bug(2) by moving special characters from valid KOI8-R places to icorrect (iso8859-1) places.