[prev in list] [next in list] [prev in thread] [next in thread] 

List:       icu-bugrfe
Subject:    [icu-bug] incoming/2497
From:       jtcsv () jtcsv ! com
Date:       2002-11-21 0:24:07
[Download RAW message or body]

schererm moved PR#2497 from incoming to others	URL: http://www.jtcsv.com/cgi-bin/icu-bugs?findid=2497

====> ORIGINAL MESSAGE FOLLOWS <====

From: weiv@jtcsv.com
Date: Thu Nov 14 21:18:04 2002
Subject: Preservation of warning codes

Full_Name: Vladimir Weinstein
Version: 2.4
OS: all
PROJECT: ICU4C
JAVA:   
Submission from: (NULL) (66.32.63.85)
Submitted by: weiv


There is an interesting thing I stumbled upon while checking return codes for
ucol_open. As one might remember, cached resource bundles remember their error
code (whether they were found, were fallback or default). On opening of a
resource bundle, the code of a cached resource is returned. This all works fine
if you come in with a U_ZERO_ERROR or an error set (if an error is set, we
promptly return as per coding conventions). However, if the passed status is set
to a warning code, it is possible that this warning code will be overwritten by
the status of the cached resource. Although this seems easy to fix, it requires
more discussion: 
- If we come with a warning code, and the resource produces a warning code, what
is the resulting warning? 
- If we came in with a warning and resource is U_ZERO_ERROR, should the warning
be overwritten?
It seems to me that this all leads to the question of how persistant are the
warning codes and which warning code should be the result. With errors, it seems
to me that the first error is preserved, as we immediately return after
encountering an error.

Let's discuss this more and make changes to the appropriate code.

_______________________________________________
icu-bugrfe mailing list
icu-bugrfe@oss.software.ibm.com
http://oss.software.ibm.com/developerworks/oss/mailman/listinfo/icu-bugrfe
[prev in list] [next in list] [prev in thread] [next in thread]