Problem with SimpleXML processing COUNTER report namespaces. #485

vinny75 · 2018-09-13T17:40:09Z

PHP Warning: simplexml_load_string(): namespace error : Namespace prefix rr on Report is not defined in /var/www/html/coral3/usage/admin/classes/domain/SushiService.php on line 611, referer: /coral3/usage/sushi.php

PHP Fatal error: Call to a member function attributes() on null in /var/www/html/coral3/usage/admin/classes/domain/SushiService.php on line 630, referer: /coral3/usage/sushi.php

Traced these errors back to:
610 $clean_xml = str_ireplace(['s:','SOAP-ENV:','SOAP:'],'',$string);
611 $xml = simplexml_load_string($clean_xml);
and
629 $report = $xml->Body->ReportResponse->Report->Report;
630 $reportTypeName = $report->attributes()->Name;

These changes seem to fix it for me
610 (line deleted)
611 $xml = simplexml_load_string($string);
612 $xml->registerXPathNamespace("s", "http://schemas.xmlsoap.org/soap/envelope/");
613 $xml->registerXPathNamespace("r", "http://www.niso.org/schemas/counter");
and
629 $report = $xml->xPath("s:Body//r:Report")[0];

vinny75 · 2018-12-10T16:38:13Z

This fails if a namespace is used within the Report element. The only reliable solution I found was to get all the namespace used in the document, then remove them.

609  $string = file_get_contents($fName);
//        $clean_xml = str_ireplace(['s:','SOAP-ENV:','SOAP:'],'',$string);
        $xml = simplexml_load_string($string);
        $namespaces = $xml->getDocNamespaces(TRUE);
        $clean_xml = $string;
        foreach ($namespaces as $namespace => $uri) {
          if (!empty($namespace)) $clean_xml = str_ireplace($namespace . ":", '', $string);
        }
        $xml = simplexml_load_string($clean_xml);

This seems like a bit of a hack but I couldn't get anything else to work.

haozeng0 · 2019-03-16T01:52:27Z

Same error, 3.0.1

PHP Fatal error: Call to a member function attributes() on null in C:\Apache24\htdocs\coral\usage\admin\classes\domain\SushiService.php on line 630

vinny75 · 2019-03-18T14:18:21Z

Same error, 3.0.1

PHP Fatal error: Call to a member function attributes() on null in C:\Apache24\htdocs\coral\usage\admin\classes\domain\SushiService.php on line 630

Did you try my second fix? If it works for you, I'll submit a pull request to get it included in the code

haozeng0 · 2019-03-18T14:42:51Z

Hi, Your second fix gives me the error:

[Mon Mar 18 10:33:17.887759 2019] [:error] [pid 5268:tid 960] [client 10.32.48.31:60819] PHP Notice: Trying to get property of non-object in C:\Apache24\htdocs\coral\usage\admin\classes\domain\SushiService.php on line 639, referer: https://archives.yu.edu/coral/usage/sushi.php
[Mon Mar 18 10:33:17.887759 2019] [:error] [pid 5268:tid 960] [client 10.32.48.31:60819] PHP Notice: Trying to get property of non-object in C:\Apache24\htdocs\coral\usage\admin\classes\domain\SushiService.php on line 639, referer: https://archives.yu.edu/coral/usage/sushi.php
[Mon Mar 18 10:33:17.887759 2019] [:error] [pid 5268:tid 960] [client 10.32.48.31:60819] PHP Fatal error: Call to a member function attributes() on null in C:\Apache24\htdocs\coral\usage\admin\classes\domain\SushiService.php on line 640, referer: https://archives.yu.edu/coral/usage/sushi.php

And my line 639: $report = $xml->Body->ReportResponse->Report->Report;

haozeng0 · 2019-03-18T16:43:55Z

The first fix works for me.
Line 611 also needs to be deleted.
Line 611: $xml = simplexml_load_string($clean_xml);

vinny75 · 2019-03-18T17:20:30Z

Hi, Your second fix gives me the error:

[Mon Mar 18 10:33:17.887759 2019] [:error] [pid 5268:tid 960] [client 10.32.48.31:60819] PHP Notice: Trying to get property of non-object in C:\Apache24\htdocs\coral\usage\admin\classes\domain\SushiService.php on line 639, referer: https://archives.yu.edu/coral/usage/sushi.php
[Mon Mar 18 10:33:17.887759 2019] [:error] [pid 5268:tid 960] [client 10.32.48.31:60819] PHP Notice: Trying to get property of non-object in C:\Apache24\htdocs\coral\usage\admin\classes\domain\SushiService.php on line 639, referer: https://archives.yu.edu/coral/usage/sushi.php
[Mon Mar 18 10:33:17.887759 2019] [:error] [pid 5268:tid 960] [client 10.32.48.31:60819] PHP Fatal error: Call to a member function attributes() on null in C:\Apache24\htdocs\coral\usage\admin\classes\domain\SushiService.php on line 640, referer: https://archives.yu.edu/coral/usage/sushi.php

And my line 639: $report = $xml->Body->ReportResponse->Report->Report;

Which vendor were you using when you got this error?

haozeng0 · 2019-03-18T17:27:40Z

EBSCO.

vinny75 · 2019-03-18T17:45:38Z

EBSCO.

I just tried EBSCO BR2 and JR1 and it worked fine with the second changes. Can't remember which vendor was failing on the first set.

haozeng0 · 2019-03-18T17:50:08Z

EBSCO. [cid:c978341e-ba18-48c7-ae44-d961dcb0b165]

…

________________________________ From: vinny75 <notifications@github.com> Sent: Monday, March 18, 2019 1:46 PM To: coral-erm/coral Cc: Hao Zeng; Comment Subject: Re: [coral-erm/coral] Problem with SimpleXML processing COUNTER report namespaces. (#485) EBSCO. I just tried EBSCO BR2 and JR1 and it worked fine with the second changes. Can't remember which vendor was failing on the first set. — You are receiving this because you commented. Reply to this email directly, view it on GitHub<#485 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ALoQzGE2awxH0ZATE0nQ8R0SB9lappYiks5vX9DvgaJpZM4Wn476>.

nkuitse · 2019-04-10T14:15:18Z

The idea of stripping strings like "s:" from a random XML document strikes me as utter insanity. For example, if the data includes a title like "Enemies: a love story" then the title will be changed to "Enemie a love story". Not to mention that a namespace prefix such as "reports:" will be changed to "report" without the colon.

The right solution is beyond me -- I don't know PHP well enough, least of all the quirks of SimpleXML -- but couldn't we at least make this bug slightly less nasty by registering some namespaces and using XPath? Something along the following lines:

$ns = array(
'xsi' => 'http://www.w3.org/2001/XMLSchema-instance',
'soapenv' => 'http://schemas.xmlsoap.org/soap/envelope/',
'counter' => 'http://www.niso.org/schemas/counter',
'sushi' => 'http://www.niso.org/schemas/sushi',
'sushicounter' => 'http://www.niso.org/schemas/sushi/counter',
);
$xml = simplexml_load_string($string);
foreach ($ns as $pfx => $uri) {
$xml->registerXPathNamespace($pfx, $uri);
}
$reports = $xml->xpath('//counter:Report');
if ($reports && $reports[0]) {
$reportTypeName = $reports[0]->attributes()->Name;
...
}

And so on?

vinny75 · 2019-04-10T15:21:10Z

The namespaces are not standard across vendors. I tried reading all the namespaces and registering them but that didn't work.

nkuitse · 2019-04-10T15:54:44Z

I agree, but I'm not trying to fix the bug, I'm just trying to mitigate the harm that it causes. The current code is just so entirely wrong -- the hack on line 610 would, for example, completely mangle a SUSHI response that happens to use a namespace prefix ending in "s" (other than the prefix "s:" itself):

$clean_xml = str_ireplace(['s:','SOAP-ENV:','SOAP:'],'',$string);

And registering the namespaces from an associative array as my code does should make it easier to add namespaces in the future. In fact, if you have a corpus of sushistore/*.xml that you're able to share, I would be happy to identify all namespaces used in them and adjust the array accordingly. You can just grep the files for : and send me that output -- I don't need anything else.

vinny75 · 2019-04-11T11:18:58Z

I tried adding the namespaces, it didn't work for me.

nkuitse · 2019-04-12T18:37:39Z

Perhaps we should just strip all namespaces from the XML, then proceed from there. You can strip namespaces like this:

function stripNamespaces($xml) {
$xsl = new DOMDocument();
$xsl->loadXML('
<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
<xsl:output method="xml" indent="no"/>
<xsl:template match="/|comment()|processing-instruction()">
xsl:copy

xsl:apply-templates/
</xsl:copy>
</xsl:template>
<xsl:template match="">
<xsl:element name="{local-name()}">

<xsl:apply-templates select="@|node()"/>
</xsl:element>
</xsl:template>
<xsl:template match="@*">
<xsl:attribute name="{local-name()}">
<xsl:value-of select="."/>
</xsl:attribute>
</xsl:template>
</xsl:stylesheet>
');
$proc = new XSLTProcessor();
$xsl = $proc->importStylesheet($xsl);
$proc->setParameter(null, "", "");
$raw = new DomDocument();
$raw->load($fName);
$doc = $proc->transformToDoc($raw);
return $doc->saveXML();
}

Then you just make a small change in lines 609-611:

$string = file_get_contents($fName);
$clean_xml = stripNamespaces($string);
$xml = simplexml_load_string($clean_xml);

Hopefully that does the trick -- I haven't tested it yet.

nkuitse · 2019-04-12T19:45:10Z

I have a sandbox instance where I can test the code in the pull request, but I have to set up a SUSHI connection in it first (and have never done that before) -- so if you have an instance you can test it in now, that would be great.

nkuitse · 2019-05-23T11:49:15Z

@queryluke has a much better fix in #579

andyp-uk · 2019-10-24T12:14:26Z

I have merged @queryluke 's fix for this issue, PR #611

nkuitse mentioned this issue Apr 12, 2019

Strip all namespaces from SUSHI responses #579

Closed

andyp-uk closed this as completed Oct 24, 2019

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Problem with SimpleXML processing COUNTER report namespaces. #485

Problem with SimpleXML processing COUNTER report namespaces. #485

vinny75 commented Sep 13, 2018

vinny75 commented Dec 10, 2018 •

edited

Loading

haozeng0 commented Mar 16, 2019

vinny75 commented Mar 18, 2019

haozeng0 commented Mar 18, 2019

haozeng0 commented Mar 18, 2019

vinny75 commented Mar 18, 2019

haozeng0 commented Mar 18, 2019

vinny75 commented Mar 18, 2019

haozeng0 commented Mar 18, 2019 via email

nkuitse commented Apr 10, 2019

vinny75 commented Apr 10, 2019

nkuitse commented Apr 10, 2019

vinny75 commented Apr 11, 2019

nkuitse commented Apr 12, 2019

nkuitse commented Apr 12, 2019

nkuitse commented May 23, 2019

andyp-uk commented Oct 24, 2019

Problem with SimpleXML processing COUNTER report namespaces. #485

Problem with SimpleXML processing COUNTER report namespaces. #485

Comments

vinny75 commented Sep 13, 2018

vinny75 commented Dec 10, 2018 • edited Loading

haozeng0 commented Mar 16, 2019

vinny75 commented Mar 18, 2019

haozeng0 commented Mar 18, 2019

haozeng0 commented Mar 18, 2019

vinny75 commented Mar 18, 2019

haozeng0 commented Mar 18, 2019

vinny75 commented Mar 18, 2019

haozeng0 commented Mar 18, 2019 via email

nkuitse commented Apr 10, 2019

vinny75 commented Apr 10, 2019

nkuitse commented Apr 10, 2019

vinny75 commented Apr 11, 2019

nkuitse commented Apr 12, 2019

nkuitse commented Apr 12, 2019

nkuitse commented May 23, 2019

andyp-uk commented Oct 24, 2019

vinny75 commented Dec 10, 2018 •

edited

Loading