|
php.net | support | documentation | report a bug | advanced search | search howto | statistics | random bug | login |
PatchesPull RequestsHistoryAllCommentsChangesGit/SVN commits
[2010-11-19 05:03 UTC] pajoye@php.net
-Status: Open
+Status: Duplicate
[2010-11-19 05:03 UTC] pajoye@php.net
|
|||||||||||||||||||||||||||
Copyright © 2001-2025 The PHP GroupAll rights reserved. |
Last updated: Mon Oct 27 05:00:01 2025 UTC |
Description: ------------ Paths in Japanese WinXP are all handled in Shift-JIS encoding ('SJIS-win' for mb_convert_encoding(), etc.). Shift-JIS contains a number of commonly-used characters for which the second byte is 0x5C (e.g. 表, 0x955C). 0x5C happens to be the encoding for backslash \, the escape character. When one of these characters is contained within a path, a large number of filesystem-related functions will fail to interpret the path correctly, and will simply not work, or will do weird things such as create or read unrequested files (see the provided test script for details). Test script: --------------- <?php // Assume file '表.txt' exists in the cwd mb_internal_encoding('SJIS'); echo filesize('表.txt')); // This works ok echo file_get_contents('表.txt'); // This fails: 'failed to open stream: No such file or directory' // Furthermore, if "表表.txt" exists then its contents will be shown by the previous command (!?!?) file_put_contents('表.txt', 'blahblahblah'); // A new file '表表.txt' is created (!?!?) ?> Expected result: ---------------- file_get_contents() should read the correct file and not fail. file_put_contents() should not create unrequested files. Actual result: -------------- file_get_contents(), is_file(), etc. fail to correctly interpret any path containing 表 and other SJIS chars with a second byte of 0x5C.