<div dir="ltr"><div><div>Hi Enio<br><br></div>Thank you, thank you. That's done the trick, and I probably wouldn't have found it myself. (And it would still have been broken after a re-install.) I've attached the diff I applied, in case anyone else needs it.<br><br></div>Rob<br></div><div class="gmail_extra"><br><div class="gmail_quote">On 22 November 2016 at 10:02, Enio Carboni <span dir="ltr"><<a href="mailto:enio.carboni@gmail.com" target="_blank">enio.carboni@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div bgcolor="#FFFFFF" text="#000000">
<p>Hi Rob,</p>
<p> i think the problem is the new version of "perl".</p>
<p>From 5.16+ the utf8::decode($param) return e different value if
$param is "undef" so you can try apply a patch
<a class="m_-5475622335681888886moz-txt-link-freetext" href="https://github.com/eprints/eprints/commit/157db0d27d423d38f9c0fc31ce47e519cd861d45" target="_blank">https://github.com/eprints/<wbr>eprints/commit/<wbr>157db0d27d423d38f9c0fc31ce47e5<wbr>19cd861d45</a></p>
<p>or use:</p>
<pre><font size="-1">sed -i -e 's/utf8::decode($value);/if( EPrints::Utils::is_set( $value ) ) {\n\t\tutf8::decode($value);\<wbr>n\t}/' perl_lib/EPrints/Repository.pm
sed -i -e 's/utf8::decode($_) for @result;/for( @result ) {\n\t\tnext if( !EPrints::Utils::is_set( $_ ) );\n\t\tutf8::decode($_);\n\t}<wbr>/' perl_lib/EPrints/Repository.pm</font><span class="HOEnZb"><font color="#888888">
</font></span></pre><span class="HOEnZb"><font color="#888888">
<br>
</font></span><div class="m_-5475622335681888886moz-cite-prefix"><span class="HOEnZb"><font color="#888888"> Enio</font></span><div><div class="h5"><br>
<br>
Il 21/11/2016 15:43, Rob McMahon ha scritto:<br>
</div></div></div>
<blockquote type="cite"><div><div class="h5">
<div dir="ltr">
<div>
<div>
<div>
<div>
<div>
<div>
<div>
<div>
<div>
<div>
<div><span id="m_-5475622335681888886gmail-goog_1963537645"></span><span id="m_-5475622335681888886gmail-goog_1963537646"></span>This
is a new installation, to move a current
archive from Ubuntu to CentOS.<br>
EPrints 3.3.10<br>
</div>
CentOS 7, with SELinux enabled<br>
<br>
</div>
I've seen a couple of references to this
problem in the archives.<br>
<a href="http://www.eprints.org/tech.php/thread-18670.html" target="_blank">[EP-tech]
EPrints 3.3.12 Installation (from source)</a>
- Gilles Fournié<br>
</div>
who refers to<br>
<a href="http://www.eprints.org/tech.php/18174.html" target="_blank">[EP-tech]
Move 302</a><font size="2"> - </font>David
McElroy<br>
</div>
But I don't see an answer, except to re-install.
I'd much rather not go down that route.<br>
<br>
</div>
I've installled 3.3.10 from source (3.3.10 because I
want the new box to look as much like the original
as possible before upgrading it). This is on CentOS
7, and I had a few issues with permissions
initially, but which I now believe are resolved,
mostly by putting the eprints directory in the
httpd_sys_content_t context.<br>
<br>
</div>
When you try to login as the admin user that I set up
during epadmin create, you are taken to a page
(cgi/users/login?login_check=<wbr>1) that says<br>
<br>
302 Moved.<br>
Moved<br>
The document has moved here.<br>
<br>
</div>
The apache ssl_access_log says<br>
<br>
[21/Nov/2016:10:43:09 +0000] "GET / HTTP/1.1" 200 7984<br>
[21/Nov/2016:10:44:03 +0000] "GET /cgi/users/home
HTTP/1.1" 302 -<br>
[21/Nov/2016:10:44:03 +0000] "GET
/cgi/users/login?target=https%<wbr>3A%2F%2F---%2Fcgi%2Fusers%<wbr>2Fhome
HTTP/1.1" 401 6119<br>
[21/Nov/2016:10:51:05 +0000] "POST /cgi/users/login
HTTP/1.1" 302 -<br>
[21/Nov/2016:10:51:06 +0000] "GET
/cgi/users/login?login_check=1 HTTP/1.1" 302 184<br>
<br>
</div>
If you press the "back" button you are back to the front
screen, and appear to be logged in successfully, although
it has an error message saying "The requested screen does
not exist, or has a fatal error: ."<br>
<br>
</div>
Has anyone got any clues as to what causes this ? Can
anyone help me debug this ? I really don't want to start
again, and have no confidence that I won't just hit the same
error again. I'm tempted to use Gilles' hack.<br>
<br>
</div>
Cheers,<br>
<br>
</div>
Rob<br>
<br>
<div>
<div>
<div>
<div><br>
</div>
</div>
</div>
</div>
</div>
<br>
<fieldset class="m_-5475622335681888886mimeAttachmentHeader"></fieldset>
<br>
</div></div><span class=""><pre>*** Options: <a class="m_-5475622335681888886moz-txt-link-freetext" href="http://mailman.ecs.soton.ac.uk/mailman/listinfo/eprints-tech" target="_blank">http://mailman.ecs.soton.ac.<wbr>uk/mailman/listinfo/eprints-<wbr>tech</a>
*** Archive: <a class="m_-5475622335681888886moz-txt-link-freetext" href="http://www.eprints.org/tech.php/" target="_blank">http://www.eprints.org/tech.<wbr>php/</a>
*** EPrints community wiki: <a class="m_-5475622335681888886moz-txt-link-freetext" href="http://wiki.eprints.org/" target="_blank">http://wiki.eprints.org/</a>
*** EPrints developers Forum: <a class="m_-5475622335681888886moz-txt-link-freetext" href="http://forum.eprints.org/" target="_blank">http://forum.eprints.org/</a>
</pre>
</span></blockquote>
<br>
</div>
<br>*** Options: <a href="http://mailman.ecs.soton.ac.uk/mailman/listinfo/eprints-tech" rel="noreferrer" target="_blank">http://mailman.ecs.soton.ac.<wbr>uk/mailman/listinfo/eprints-<wbr>tech</a><br>
*** Archive: <a href="http://www.eprints.org/tech.php/" rel="noreferrer" target="_blank">http://www.eprints.org/tech.<wbr>php/</a><br>
*** EPrints community wiki: <a href="http://wiki.eprints.org/" rel="noreferrer" target="_blank">http://wiki.eprints.org/</a><br>
*** EPrints developers Forum: <a href="http://forum.eprints.org/" rel="noreferrer" target="_blank">http://forum.eprints.org/</a><br>
<br></blockquote></div><br></div>