fopen creates files with utf characters under certain conditions
Nobody has claimed this yet.
- Dominant language
- C
- Stars
- 40.4k
- Forks
- 8.1k
- Avg merge
- 2d 13h
- Merged PRs (30d)
- 96
Description
Description
The following code:
if (@$_GET["charset"]) echo("set charset from ".ini_set("default_charset", "Windows-1251")."<BR>\n");
//ini_set("default_charset", "Windows-1251");
$FN="C:/Temp/АБВГДЕЖЗ.txt"; //no UTF here!
echo("filename: ".$FN."<BR>\n");
$hFile=fopen($FN, 'w'); fwrite($hFile, $FN); fclose($hFile);
Resulted in this output: the created file was C:\TEMP\АБВГДЕЖЗ.txt
(filename got converted to Utf-8)
But I expected this output instead: the file should be C:\TEMP\АБВГДЕЖЗ.txt
Explanation:
"АБВГДЕЖЗ" here are the first letters of Russian abc. They could be any other Russian letters or non-latin letters of another non-English language / charset.
The initial setting of default_charset for me is UTF-8 (as returned by ini_set()).
This is not a trivial bug and it's not easy to reproduce it, yet it is possible.
It only happens if a script calls ini_set("default_charset", "Windows-1251");
It is related to Apache's keep-alive connections reused for later requests.
Steps to reproduce:
- open 2 tabs in your browser, one with the script as is (for ex. dev.php) and another with charset parameter (dev.php?charset=1). If you run both tabs now they both will work correct.
- restart Apache
- refresh the 1st tab first several times (at least 2, the more the better). This creates Apache threads with a context of Utf-8. The created file is Ok.
- refresh the 2nd tab once. It creates a filename with Utf-8 characters. The reason is it is processed by a thread that just served the same script without changing default_charset.
For some time, no matter how many times you refresh 2nd tab, the file created will have Utf chars. But after some time the phenomenon will stop. Also, the bug may show even without restarting Apache, if you first call 1st tab many times, then wait for some time, like 2-3 minutes. But it takes longer.
This bug affects other files-related functions like file_exists() etc.
Both Php and Apache have been upgraded to their last versions: Apache/2.4.58 (Win64) PHP/8.3.3
PHP Version
8.3.3
Operating System
Windows 8.1
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with the supplied reproduction script on PHP 8.3.3 under Apache on Windows, exercising fopen(), file_exists(), default_charset, and keep-alive requests in the stated tab sequence. Trace the file-related behavior across reused Apache threads and compare requests before and after changing default_charset. Done means non-Latin filenames remain consistent regardless of request order or connection reuse.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- apache, php
- Domain
- backend, operating-systems
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100