Sortie d’apache 1.3.28

Notre server web prefere a tous soit Apache vient de sortir
une nouvelle version dans la serie 1.3 soit la 1.3.28.

Annonce de la sortie

Au menu : Rien de tres nouveau

Security vulnerabilities

The main security vulnerabilities addressed in 1.3.28 are:

* CAN-2003-0460 (cve.mitre.org): Fix the rotatelogs support program on Win32 and OS/2 to ignore special control characters received over the pipe. Previously such characters could cause it to quit logging and exit. We would like to thank the Hitachi Incident Response team for their responsible disclosure of this issue.
* VU#379828 : The server could crash when going into an infinite loop due to too many subsequent internal redirects and nested subrequests.
* Eliminated leaks of several file descriptors to child processes, such as CGI scripts.

New features

The main new features in 1.3.28 (compared to 1.3.27) are:

* Added new ap_register_cleanup_ex() API function which allows for a “magic” cleanup function to be run at register time rather than at cleanup time.
* Improvements to mod_usertrack that allows for a regular (verbose) as well as “compact” version of the tracking cookie (the new ‘CookieFormat’ directive), and the ability to prepend a string to the cookie via the ‘CookiePrefix’ directive.

New features that relate to specific platforms:

* Introduce Win32 .pdb diagnostic symbols into the Apache 1.3 build (as created in Apache 2.0.45 and later.) which makes debugging and analysis of crash dumps and Dr. Watson logs trivial.
* AIX: Change the default accept mutex mechanism from pthread back to fcntl.

Bugs fixed

The following bugs were found in Apache 1.3.27 (or earlier) and have been fixed in Apache 1.3.28:

* Make sure the accept mutex is released before calling child exit hooks and cleanups.
* Fix mod_rewrite’s handling of absolute URIs. The escaping routines now work scheme dependent and the query string will only be appended if supported by the particular scheme.
* Prevent obscenely large values of precision in ap_vformatter from clobbering a buffer.
* Update timeout algorithm in free_proc_chain. If a subprocess did not exit immediately, the thread would sleep for 3 seconds before checking the subprocess exit status again. In a very common case when the subprocess was an HTTP server CGI script, the CGI script actually exited a fraction of a second into the 3 second sleep, which effectively limited the server to serving one CGI request every 3 seconds across a persistent connection.

Bonne Compilation…

Be Sociable, Share!

•