gistfile1.txt

<?xml version="1.0" standalone="yes" ?>
<!DOCTYPE bugzilla SYSTEM "http://bugs.bricolage.cc/bugzilla.dtd">

<bugzilla version="3.0"
          urlbase="http://bugs.bricolage.cc/"
          maintainer="[email protected]"
          exporter="[email protected]"
>

    <bug>
          <bug_id>802</bug_id>
          
          <creation_ts>2004-09-07 19:17</creation_ts>
          <short_desc>Set SvUTF8_on on Data Fetched from Database</short_desc>
          <delta_ts>2004-09-10 23:10:47</delta_ts>
          <product>Bricolage</product>
          <bug_status>ASSIGNED</bug_status>
          
          <priority>P2</priority>
          <bug_severity>enhancement</bug_severity>
          <target_milestone>1.8.3</target_milestone>
          
          <everconfirmed>1</everconfirmed>
          <reporter name="David Wheeler">[email protected]</reporter>
          <assigned_to name="David Wheeler">[email protected]</assigned_to>
          

      

      
          <long_desc isprivate="0">
            <who name="David Wheeler">[email protected]</who>
            <bug_when>2004-09-07 19:17:14</bug_when>
            <thetext>Once a new version of DBI comes out with the necessary UTF-8 tools, use

  $dbh-&gt;{SetUTF8} = 2;

To make all of our character issues go away.

References:

  http://bugs.bricolage.cc/show_bug.cgi?id=709#c14
  http://www.mail-archive.com/[email protected]/msg03451.html</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who name="David Wheeler">[email protected]</who>
            <bug_when>2004-09-09 14:32:51</bug_when>
            <thetext>Created an attachment (id=135)
Interim Solution

This is an interim solution using DBD::Pg&apos;s &quot;pg_enable_utf8&quot; attribute. This
should turn on the utf-8 flag for all data fetched from the database, thus
allowing substr() and friends to work properly. I&apos;d like to get rid of it and
use the DBI solution once it&apos;s in DBI. I&apos;m hoping that Tim can get it in for
1.44.</thetext>
          </long_desc>
          <long_desc isprivate="0">
            <who name="David Wheeler">[email protected]</who>
            <bug_when>2004-09-10 03:08:35</bug_when>
            <thetext>It turns out that the attached patch wasn&apos;t enough of an interim solution. 
With the fix for Bug # 648 committed, other things started to go haywire. 
It seems that Perl doesn&apos;t like to concatenate variables with their UTF-8 
flag with others without their UTF-8 flag set. So we were left with two 
choices:

1. Turn off UTF-8 in the data submitted from the browser (the original fix 
for Bug # 648) and then eliminate all uses of substr().

2. Turn on the UTF-8 flag everywhere so that one is always concatenating 
UTF-8 to UTF-8 and Perl is happy.

The second option turned out to be the only real option, although I did try 
for a while to do the first. It just wasn&apos;t practical. In turning on utf8 
everywhere, though, I had to do quite a bit of experimentation to get it 
right. Thanks to gugod for his help in testing!

I&apos;m going to keep this bug report open as an enhancement for now, however, 
because I still want to switch to using DBI instead of DBD::Pg for setting 
the UTF-8 flag on all data pulled from the database, just as soon as Tim is 
able to add it and get a release out.</thetext>
          </long_desc>
      
    </bug>

</bugzilla>