Many zoneminder systems have a "Can't stat : No such file or directory" in their logs each time zmfix is run. The problem was traced back to the sql query, which looked for a ControlDevice field that was not null. However, the ControlDevice field is empty, rather than null, when there is no control device.
Updates the preg_match expression to match Internet Explorer 10. The previous pattern would not detect IE10, and hence zoneminder would not attempt to stream via Cambozola, which leaves IE10 users looking at a broken link.
NOTE: There is another file in the zoneminder source tree, AssetDispatcher.php, that also uses the same pattern match expression. It is part of the Cake folder, which I am not familiar with. Please advise if we should edit that file as well.
zms crashes with hashed authentication - fixes#88
Previously, systems without gnutls were computing auth_md5 to be twice the size of what it was defined to be, thus causing zms to crash. The for loop at line 227 builds auth_md5 by looping every "2j", which means the upper limit (md5len) should be half the desired size of auth_md5.
Previously, systems without gnutls were computing auth_md5 to be twice the size of what it was defined to be, thus causing zms to crash. The for loop at line 227 builds auth_md5 by looping every "2j", which means the upper limit (md5len) should be half the desired size of auth_md5.
Noticed the init script was looking for a zm_update, which doesn't exist. My best guess is that this should be changed to zmupdate.pl. This causes zoneminder to reload its configuration from the database prior to startup. Thoughts?
When upgrading the zoneminder database to 1.26, this proposed change will ask the user to convert existing MyISAM tables over InnoDB. It first scans for existing MyISAM tables and will skip over any tables that are already converted to InnoDB.
I've tested this on (a copy of) a production database server and have verified it doesn't alter existing tables from a different schema.
And I'm done messing with the zmupdate script. I promise.
I understand if this change is considered outside the scope of a current bug fix release.
Added frameserver patch from Zoneminder Wiki
Changes the socket reader in zmf from a single read, to a loop read.
Incomplete reads would be reported even though the image writer wrote
the whole image to the socket. The problem was when the read went to
read the image frame from the socket, all the data had not yet been written
to the socket by the writer, so the reader thought there was a problem.
The loop reads from the socket until a full image frame is read, or there is
an error.
Originally at http://www.zoneminder.com/wiki/index.php/1.24.2_Patches
I have long ago converted my production zoneminder dB tables to InnoDB and have not experienced any side effects. One benefit of using InnoDB is a significantly lesser chance of table corruption after an unclean shutdown. Thoughts?
Isaac Connor
Kyle Johnson
Knniggett
Chris Wiggins
Andrew Bauer