Received: (qmail 17764 invoked by uid 2012); 6 Jul 1998 10:14:38 -0000 Message-Id: <19980706101438.17763.qmail@hyperreal.org> Date: 6 Jul 1998 10:14:38 -0000 From: Dennis Lundberg Reply-To: dennis.lundberg@mdh.se To: apbugs@hyperreal.org Subject: URL:s containing the character 'ö' gets trucated (See also PR 800) X-Send-Pr-Version: 3.2 >Number: 2553 >Category: general >Synopsis: URL:s containing the character 'ö' gets trucated (See also PR 800) >Confidential: no >Severity: serious >Priority: medium >Responsible: apache >State: closed >Class: sw-bug >Submitter-Id: apache >Arrival-Date: Mon Jul 6 03:20:00 PDT 1998 >Last-Modified: Mon Jul 20 20:39:55 PDT 1998 >Originator: dennis.lundberg@mdh.se >Organization: >Release: 1.3.0 >Environment: uname -a: SunOS aragorn 5.5.1 Generic_103640-01 sun4u sparc SUNW,Ultra-1 Compiler: gcc 2.8.1 >Description: This problem was reported earlier as PR 800 for Apache 1.2.x and was 'mysteriously' fixed in the 1.3-beta source tree. The problem did not exist in 1.3b5 but reappeared in 1.3.0. When a URL includes the character 'ö', Apache truncates the URL at the position where the 'ö' character is, and tries to serve that URL to the client. Example: http://www.company.com/someöfile.htm is trucated by Apache 1.3.0 to: http://www.company.com/some Other international character like å, ä, Å, Ä and Ö work just fine. >How-To-Repeat: Create a file named 'fileö.htm' and try to access from a web running Apache 1.3.0. I'm sorry for not beeing able to supply a URL for this but our webserver depends on files containing the character 'ö'. Therefor I have now downgraded to 1.3b5 as a work-around. >Fix: Something must have changed between 1.3b5 and 1.3.0 in the source. I'm not qualified enough to guess where this change might be though. >Audit-Trail: State-Changed-From-To: open-closed State-Changed-By: coar State-Changed-When: Mon Jul 6 04:26:31 PDT 1998 State-Changed-Why: That character is not legal verbatim in URLs. From RFC1738: : No corresponding graphic US-ASCII: : : URLs are written only with the graphic printable characters of the : US-ASCII coded character set. The octets 80-FF hexadecimal are not : used in US-ASCII, and the octets 00-1F and 7F hexadecimal represent : control characters; these must be encoded. The ö character is ISO-Latin-1 position 246(d), which is outside the range. Therefore, this character needs to be encoded in the URL as %F6. Thanks for using Apache! State-Changed-From-To: closed-analyzed State-Changed-By: dgaudet State-Changed-When: Mon Jul 6 21:17:51 PDT 1998 State-Changed-Why: Until someone tracks this down on solaris it should still be open. Dean Comment-Added-By: dgaudet Comment-Added-When: Mon Jul 6 21:50:15 PDT 1998 Comment-Added: Try adding "EXTRA_CFLAGS=-funsigned-char" to your Configuration, reconfiguring and recompiling. This is a workaround until we cleanup the use of the isspace()/etc. macros. Dean From: Dennis Lundberg To: dgaudet@apache.org Cc: apbugs@apache.org Subject: Re: general/2553: URL:s containing the character =?iso-8859-1?Q?'=F6'?= gets trucated (See also PR 800) Date: Tue, 07 Jul 1998 10:23:28 +0200 Adding "EXTRA_CFLAGS=-funsigned-char" to my Configuration, reconfiguring and recompiling solved the problem for me on Solaris. Thanks! /Dennis State-Changed-From-To: analyzed-closed State-Changed-By: dgaudet State-Changed-When: Mon Jul 20 20:39:55 PDT 1998 State-Changed-Why: fixed in 1.3.1 >Unformatted: [In order for any reply to be added to the PR database, ] [you need to include in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ]