php.net |  support |  documentation |  report a bug |  advanced search |  search howto |  statistics |  random bug |  login
Bug #34832 sprintf not reentrant
Submitted: 2005-10-11 17:31 UTC Modified: 2007-01-26 01:00 UTC
Votes:12
Avg. Score:5.0 ± 0.0
Reproduced:12 of 12 (100.0%)
Same Version:10 (83.3%)
Same OS:11 (91.7%)
From: jdolecek at NetBSD dot org Assigned:
Status: No Feedback Package: Strings related
PHP Version: 5CVS-2005-10-17 (snap) OS: Windows 2000
Private report: No CVE-ID: None
View Add Comment Developer Edit
Welcome! If you don't have a Git account, you can't do anything here.
You can add a comment by following this link or if you reported this bug, you can edit this bug over here.
(description)
Block user comment
Status: Assign to:
Package:
Bug Type:
Summary:
From: jdolecek at NetBSD dot org
New email:
PHP Version: OS:

 

 [2005-10-11 17:31 UTC] jdolecek at NetBSD dot org
Description:
------------
I'm using PHP with Apache 1.3.23 via apache module, using windows PHP binaries downloaded from php.net.

If several parallelly running scripts execute same blocks of code using sprintf("%.2f", $somefloat*$otherfloat), occassionally incorrect float number is printed or even one character of the "float" contains binary NUL.

Easy way to trigger this is to have a FPDF script generating a PDF document, and have something which calls this script via web request in parallel, so that all scripts are executed within the context of single web server process. The problem appears in about 1% of generated documents. The problem is repeatably in my environment, I can try to put together simplier version of reproducing script if need be.

If I replace the above contruct with number_format($somefloat*$otherfloat, 2, '.', ''), it works without problem and no problem occurs.

The length of the gardbled output always exactly matches the length of non-gardbled output, FWIW.

This suggests there is some kind of reentrancy problem with the sprintf() routine, at least when printing floats. Perhaps some stack alignment issue or something equally esoteric?

The problem exists in 5.1b3, 5.1.0rc1, 5.0.5. I did not check other versions.

The type of bug has been set to repoducible crash due to missing 'Memory corruption' category and since this _might_ cause a crash (tho no crash has been observed) - feel free to change to anything more appropriate if need be.

Bug #21659 might actually be same problem, tho #21659 has been shrugged off as miscompilation problem.



Patches

Add a Patch

Pull Requests

Add a Pull Request

History

AllCommentsChangesGit/SVN commitsRelated reports
 [2005-10-12 12:07 UTC] dolecek at stringdata dot cz
Reproduce steps:
1. download FPDF 1.53 (fpdf153.zip), extract to suitable directory accessible via web server
2. edit fpdf.php and replace line:

$this->_out('/CreationDate '.$this->_textstring('D:'.date('YmdHis')));

with:

$this->_out('/CreationDate '.$this->_textstring('D:20051011143910'));

This is necessary so that generated PDF don't differ gratuitously.

2. put attached script to the same directory as the modified fpdf.php

3. run:
      siege -c 10 -v -u http://server/path/to/script.php

The script:

$pdf=new FPDF();
$pdf->AddPage();

$pdf->SetFont('Arial','B',16);
$pdf->Cell(40,10,'Hello World!');
$pdf->SetFont('Arial','',10);
$pdf->Cell(60,10,'Powered by FPDF.',0,1,'C');

$content = $pdf->Output(false, 'S');

if (md5($content) != '38ec5b17d2cda871e21b28ecaeecd08d') {
   trigger_error('pdf generation test bug', E_USER_ERROR);
}
 [2005-10-17 09:56 UTC] dolecek at stringdata dot cz
Tested with php5.1-win32-200510170630 and php5.0-win32-200510170030.

PHP 5.1 crashes Apache after several seconds of running the test. PHP 5.0.6-dev doesn't crash, but still exhibits the problem. Thus, the problem is not fixed in snapshot.
 [2005-11-01 11:43 UTC] sniper@php.net
Thank you for this bug report. To properly diagnose the problem, we
need a short but complete example script to be able to reproduce
this bug ourselves. 

A proper reproducing script starts with <?php and ends with ?>,
is max. 10-20 lines long and does not require any external 
resources such as databases, etc.

If possible, make the script source available online and provide
an URL to it here. Try to avoid embedding huge scripts into the report.


 [2005-11-01 17:15 UTC] dolecek at stringdata dot cz
Reproducing script already provided in previous feedback.
 [2005-11-02 13:26 UTC] sniper@php.net
Yes, and now please read the comment. I asked for short but complete script..and not some huge package.
 [2005-11-10 01:00 UTC] php-bugs at lists dot php dot net
No feedback was provided for this bug for over a week, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".
 [2007-01-18 09:01 UTC] jdolecek at NetBSD dot org
(non content related, just editing submissing e-mail)
 [2007-01-18 09:11 UTC] tony2001@php.net
Please try using this CVS snapshot:

  http://snaps.php.net/php5.2-latest.tar.gz
 
For Windows:
 
  http://snaps.php.net/win32/php5.2-win32-latest.zip


 [2007-01-26 01:00 UTC] php-bugs at lists dot php dot net
No feedback was provided for this bug for over a week, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".
 
PHP Copyright © 2001-2021 The PHP Group
All rights reserved.
Last updated: Sun Nov 28 05:03:12 2021 UTC