www/172195: PR database corrupts patches
Michael Gmelin
freebsd at grem.de
Sun Sep 30 21:40:05 UTC 2012
>Number: 172195
>Category: www
>Synopsis: PR database corrupts patches
>Confidential: no
>Severity: non-critical
>Priority: low
>Responsible: freebsd-www
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Sun Sep 30 21:40:04 UTC 2012
>Closed-Date:
>Last-Modified:
>Originator: Michael Gmelin
>Release: FreeBSD 9.0
>Organization:
Grem Equity GmbH
>Environment:
FreeBSD undermydesk 9.0-RELEASE FreeBSD 9.0-RELEASE #0: Tue Jan 3 07:46:30 UTC 2012
>Description:
Uploading patches to the FreeBSD PR system might corrupt them on download from the PR web page. This affects UTF-8 encoded patches as well as binary patches.
As an additional test I will upload a text containing a UTF-8 snippet, which should resemble two 3 byte uTF-8 sequences forming the word China in Chinese
(0xe4 0xb8 0xad 0xe5 0x9b 0xbd or 中国)
Then I will send a follow up email, that contains the same man page as a mime attachment using content-type application/x-gzip.
I might to additional tests on this PR (more UTF-8 based text etc.) to see if there is a safe way to provide this kind of data to a PR.
>How-To-Repeat:
Send a bug follow-up containing a gzipped file or UTF-8 file containing special characters.
>Fix:
For binary patches: Copy and paste the base64 representation shown in the PR and decode it on the command line, e.g. using "openssl enc -d -a".
For text: none known.
This might not be a bug and completely work as designed - in this case I would like to see some documentation though, since these days patches might contain UTF-8 characters on a regular basis.
Patch attached with submission follows:
ä¸å½
>Release-Note:
>Audit-Trail:
>Unformatted:
More information about the freebsd-www
mailing list