Ticket #1552 (closed enhancement: duplicate)

SVN Diffs for #1552

 

Opened 7 years ago

Last modified 1 year ago

Default converter determination should recognize non-standard names

Reported by: markus.scherer(at)jtcsv.com Assigned to: grhoten
Priority: trivial Milestone: UNSCH
Component: conversion Version: 2.0
Keywords: Cc:
Load: Xref: 3449
Java Version: Operating System: all
Project (C/J): all Weeks: 0.2
Review:

Description (Last modified by grhoten)

We keep getting reports that ICU can not open a proper default converter on certain platforms even if it has the actually necessary converter.

This happens mainly with non-standard names (not IANA/MIME-registered etc.) like: Solaris - 646 (US-ASCII), 5601 (KSC_5601) Linux - ujis (EUC-JP?), eucjputf8 (UTF-8?!)

Also seen on Linux are charset names that by themselves do not specify a charset at all. Particularly, Linux locale IDs japanese.euc - should be ja_JP.EUC-JP korean.euc - should be ko_KR.EUC-KR chinese.euc - should be zh_CN.EUC-CN(?)(is there one for "taiwanese"/EUC-TW??)

Note that the above are not even valid locale IDs (these are not valid language codes)!

Instead of polluting our alias table (convrtrs.txt) with such bad names, we should add additional names lists and smart logic to the default codepage determination in putil.c.

Attachments

Change History

12/31/69 17:32:05 changed by notes2

New UTF-8 locales on various operating systems have lowered demand for this feature [grhoten 2004-01-31]

12/31/69 17:32:06 changed by notes

This should be fixed in ICU4C 3.4.1 and later.

12/31/69 17:32:07 changed by auditor

  • 11/27/01 12:40:13 schererm changed notes
  • 11/27/01 12:40:14 schererm moved from incoming to conversion
  • 07/29/02 18:54:48 schererm changed notes
  • 10/29/02 15:07:29 hshih changed notes2
  • 05/29/03 16:04:55 hshih changed notes2
  • 02/01/04 02:02:35 grhoten changed notes2
  • 02/09/04 15:22:16 grhoten changed notes2
  • 07/09/04 15:03:04 grhoten changed notes2
  • 07/12/04 17:36:22 mark changed notes2
  • 07/16/04 18:28:05 grhoten changed notes2
  • Fri Dec 2 16:05:57 2005 grhoten changed notes2: xref: "" to "3449",
  • Wed Feb 22 11:19:39 2006 guest sent reply 1
  • Fri Oct 13 23:24:18 2006 grhoten changed notes2: target: "UNSCH" to "3.8 candidate",
  • Fri Oct 13 23:24:18 2006 grhoten changed notes
  • Wed Nov 22 07:44:26 2006 grhoten changed notes2: target: "3.8 candidate" to "UNSCH",
  • Mon Nov 27 17:25:03 2006 schererm changed notes2: summmary: "" to "RFE: default converter determination should recognize non-standard names",

02/22/06 10:19:39 changed by chornig(at)us.ibm.com

(Guest Reply)

Some more I've run into are "ko_KR.EUC" and "zh_TW.EUC" on Solaris.

10/03/07 21:59:49 changed by grhoten

  • load changed.
  • status changed from new to closed.
  • java changed.
  • description changed.
  • summary changed from RFE: default converter determination should recognize non-standard names to Default converter determination should recognize non-standard names.
  • project set to all.
  • keywords deleted.
  • resolution set to duplicate.
  • revw changed.

These issues have been resolved with #3449

Add/Change #1552 (Default converter determination should recognize non-standard names)




Anti spam check: