php.net |  support |  documentation |  report a bug |  advanced search |  search howto |  statistics |  random bug |  login
Bug #71630 copy function does not work if destination file does not exist
Submitted: 2016-02-19 12:39 UTC Modified: 2016-02-19 17:49 UTC
Votes:1
Avg. Score:5.0 ± 0.0
Reproduced:0 of 0 (0.0%)
From: vladimir at volomp dot com Assigned: cmb (profile)
Status: Not a bug Package: *General Issues
PHP Version: 5.5.32 OS: linux/all
Private report: No CVE-ID: None
View Developer Edit
Welcome! If you don't have a Git account, you can't do anything here.
If you reported this bug, you can edit this bug over here.
(description)
Block user comment
Status: Assign to:
Package:
Bug Type:
Summary:
From: vladimir at volomp dot com
New email:
PHP Version: OS:

 

 [2016-02-19 12:39 UTC] vladimir at volomp dot com
Description:
------------
Fails to copy a.txt to b.txt if b.txt is missing.
If we just touch b.txt, copy will work.


There is possibly relevant code in ext/standard/file.c in function php_copy_file_ctx() where dest is checked the same way source is (suspected copy&paste).

        switch (php_stream_stat_path_ex(dest, PHP_STREAM_URL_STAT_QUIET | PHP_STREAM_URL_STAT_NOCACHE, &dest_s, ctx)) {
                case -1:
                        /* non-statable stream */
                        goto safe_to_copy;
                        break;
                case 0:
                        break;
                default: /* failed to stat file, does not exist? */
                        return ret;
        }


So it is unclear if PHP_STREAM_URL_STAT_QUIET has any influence on return code - but I guess NO, so then this block of code is incorrect and should have default goto safe_to_copy.

This code is more-less same for php 5.6 and 7, however 5.2.17 which we still use do not have this bug as check is done differently:

        if (php_stream_stat_path_ex(dest, PHP_STREAM_URL_STAT_QUIET, &dest_s, NULL) != 0) {
                goto safe_to_copy;
        }




Test script:
---------------
copy( "a.txt", "b.txt" );




Patches

php_copy_file_ctx-destination-check-fix (last revision 2016-02-19 12:40 UTC by vladimir at volomp dot com)

Pull Requests

History

AllCommentsChangesGit/SVN commitsRelated reports
 [2016-02-19 13:28 UTC] danack@php.net
-Status: Open +Status: Feedback
 [2016-02-19 13:28 UTC] danack@php.net
Please can you provide some context of how you are encountering a problem. i.e. provide a reproduce scripts.


The code below works for me:

<?php

file_put_contents("a.txt", "These are some words");
copy("a.txt", "b.txt");
var_dump(file_exists("b.txt"));

//output is bool(true)

without b.txt existing first.
 [2016-02-19 13:38 UTC] vladimir at volomp dot com
-Status: Feedback +Status: Open
 [2016-02-19 13:38 UTC] vladimir at volomp dot com
Your script is just fine for testing. I get bool(false) consistently.

php --version

PHP 5.5.16 (cli) (built: Aug 31 2014 16:22:37)
Copyright (c) 1997-2014 The PHP Group
Zend Engine v2.5.0, Copyright (c) 1998-2014 Zend Technologies
    with Zend OPcache v7.0.4-dev, Copyright (c) 1999-2014, by Zend Technologies


<?php

file_put_contents("a.txt", "These are some words");
copy("a.txt", "b.txt");
var_dump(file_exists("b.txt"));



I will git clone and check completely stock .32
 [2016-02-19 13:48 UTC] danack@php.net
Please can you confirm that you don't have any security modules active that would prevent creation of new files - such as SELinux or Suhosin?

I think it is far more likely that it is some problem unique to your computer rather than a general bug with copy().
 [2016-02-19 13:56 UTC] vladimir at volomp dot com
I have patched my server with same check as in 5.2.17 and it now works flawlessly. Could somebody explain how could this actually work? 

Is it php_stream_stat_path_ex() supposed to return 0 if file is not found?  This code just doesn't seem right.
 [2016-02-19 14:03 UTC] vladimir at volomp dot com
Seems that git 5.5.32 works just fine, as expected. I do have other patches inside, as this is redhat based rpm. Sorry, off I go to check further. But still I do not understand how could it possibly work with this switch/case?

When file does not exist, should -1 as 'non-statable stream' be expected, or some other error code?
 [2016-02-19 14:28 UTC] danack@php.net
-Status: Open +Status: Feedback
 [2016-02-19 14:28 UTC] danack@php.net
> When file does not exist, should -1 as 'non-statable stream' be expected,

I believe that is correct.

> Is it php_stream_stat_path_ex() supposed to return 0 if file is not found?

I believe the return values mirror those of standard system stat calls - 0 indicates success, other values a errors. http://linux.die.net/man/2/stat


If you're not seeing the error on git 5.5.32, then you'll need to isolate where the error has been introduced as we are unable to reproduce it.
 [2016-02-19 15:40 UTC] vladimir at volomp dot com
-Status: Feedback +Status: Open
 [2016-02-19 15:40 UTC] vladimir at volomp dot com
Danack, thanks for your input. I have solved the problem - it was indeed one of the custom patches that was faulty, causing stat to return a bit unexpected value 2, that came from errno by mistake.
 [2016-02-19 17:49 UTC] cmb@php.net
-Status: Open +Status: Not a bug -Assigned To: +Assigned To: cmb
 [2016-02-19 17:49 UTC] cmb@php.net
So closing as "not a bug".
 
PHP Copyright © 2001-2024 The PHP Group
All rights reserved.
Last updated: Fri Dec 27 03:01:28 2024 UTC