|
|
@@ -1,5 +1,5 @@
|
|
|
<?php
|
|
|
-// Copyright (C) 2010-2013 Combodo SARL
|
|
|
+// Copyright (C) 2010-2014 Combodo SARL
|
|
|
//
|
|
|
// This file is part of iTop.
|
|
|
//
|
|
|
@@ -68,6 +68,7 @@ class WebPage implements Page
|
|
|
protected $sContentType;
|
|
|
protected $sContentDisposition;
|
|
|
protected $sContentFileName;
|
|
|
+ protected $bTrashUnexpectedOutput;
|
|
|
protected $s_sOutputFormat;
|
|
|
protected $a_OutputOptions;
|
|
|
|
|
|
@@ -88,6 +89,7 @@ class WebPage implements Page
|
|
|
$this->sContentType = '';
|
|
|
$this->sContentDisposition = '';
|
|
|
$this->sContentFileName = '';
|
|
|
+ $this->bTrashUnexpectedOutput = false;
|
|
|
$this->s_OutputFormat = utils::ReadParam('output_format', 'html');
|
|
|
$this->a_OutputOptions = array();
|
|
|
ob_start(); // Start capturing the output
|
|
|
@@ -392,22 +394,55 @@ class WebPage implements Page
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
- * Discard unexpected output data
|
|
|
+ * Discard unexpected output data (such as PHP warnings)
|
|
|
* This is a MUST when the Page output is DATA (download of a document, download CSV export, download ...)
|
|
|
*/
|
|
|
public function TrashUnexpectedOutput()
|
|
|
{
|
|
|
- // This protection is redundant with a protection implemented in MetaModel::IncludeModule
|
|
|
- // which detects such issues while loading module files
|
|
|
- // Here, the purpose is to detect and discard characters produced by the code execution (echo)
|
|
|
- $sPreviousContent = ob_get_clean();
|
|
|
- if (trim($sPreviousContent) != '')
|
|
|
+ $this->bTrashUnexpectedOutput = true;
|
|
|
+ }
|
|
|
+
|
|
|
+ /**
|
|
|
+ * Read the output buffer and deal with its contents:
|
|
|
+ * - trash unexpected output if the flag has been set
|
|
|
+ * - report unexpected behaviors such as the output buffering being stopped
|
|
|
+ *
|
|
|
+ * Possible improvement: I've noticed that several output buffers are stacked,
|
|
|
+ * if they are not empty, the output will be corrupted. The solution would
|
|
|
+ * consist in unstacking all of them (and concatenate the contents).
|
|
|
+ */
|
|
|
+ protected function ob_get_clean_safe()
|
|
|
+ {
|
|
|
+ $sOutput = ob_get_contents();
|
|
|
+ if ($sOutput === false)
|
|
|
{
|
|
|
- if (Utils::GetConfig() && Utils::GetConfig()->Get('debug_report_spurious_chars'))
|
|
|
+ $sMsg = "Design/integration issue: No output buffer. Some piece of code has called ob_get_clean() or ob_end_clean() without calling ob_start()";
|
|
|
+ if ($this->bTrashUnexpectedOutput)
|
|
|
{
|
|
|
- IssueLog::Error("Output already started before downloading file:\nContent was:'$sPreviousContent'\n");
|
|
|
+ IssueLog::Error($sMsg);
|
|
|
+ $sOutput = '';
|
|
|
+ }
|
|
|
+ else
|
|
|
+ {
|
|
|
+ $sOutput = $sMsg;
|
|
|
}
|
|
|
}
|
|
|
+ else
|
|
|
+ {
|
|
|
+ ob_end_clean(); // on some versions of PHP doing so when the output buffering is stopped can cause a notice
|
|
|
+ if ($this->bTrashUnexpectedOutput)
|
|
|
+ {
|
|
|
+ if (trim($sOutput) != '')
|
|
|
+ {
|
|
|
+ if (Utils::GetConfig() && Utils::GetConfig()->Get('debug_report_spurious_chars'))
|
|
|
+ {
|
|
|
+ IssueLog::Error("Trashing unexpected output:'$s_captured_output'\n");
|
|
|
+ }
|
|
|
+ }
|
|
|
+ $sOutput = '';
|
|
|
+ }
|
|
|
+ }
|
|
|
+ return $sOutput;
|
|
|
}
|
|
|
|
|
|
/**
|
|
|
@@ -419,8 +454,8 @@ class WebPage implements Page
|
|
|
{
|
|
|
header($s_header);
|
|
|
}
|
|
|
- $s_captured_output = ob_get_contents();
|
|
|
- ob_end_clean();
|
|
|
+
|
|
|
+ $s_captured_output = $this->ob_get_clean_safe();
|
|
|
echo "<!DOCTYPE html PUBLIC \"-//W3C//DTD XHTML 1.0 Strict//EN\" \"http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd\">\n";
|
|
|
echo "<html>\n";
|
|
|
echo "<head>\n";
|
romainq