- Feb 17, 2015
-
-
Jared Hancock authored
-
Jared Hancock authored
-
- Feb 13, 2015
-
-
Jared Hancock authored
-
- Feb 11, 2015
-
-
Jared Hancock authored
-
Jared Hancock authored
This patch fixes a vulnerable scenario, where sequential login attempts can be made without an existing session, and without a valid CSRF token. This scenario lends itself well for brute force password attempts, because attackers can avoid using a session and still send requests to determine if a set of credentials are valid. This vector also avoids the authentication lockout mechanism, because it requires an ongoing session to shutdown the requests. This patch addresses the issue by requiring a session and a valid CSRF token generated by the server and placed in the session to be submitted with the credentials. Therefore, an existing session and a Cookie header are required to process a login attempt. Secondly, the CSRF token will be changed on the server after each login processed. Therefore, for each session, a subsequent GET request would be necessary before submitting another login attempt.
-
- Feb 06, 2015
-
-
Jared Hancock authored
-
Jared Hancock authored
-
Jared Hancock authored
-
- Feb 03, 2015
-
-
Jared Hancock authored
References: https://bugs.php.net/bug.php?id=43200 http://stackoverflow.com/a/22521203
-
Jared Hancock authored
-
- Feb 02, 2015
-
-
Jared Hancock authored
-
- Jan 30, 2015
-
-
Jared Hancock authored
-
- Jan 24, 2015
-
-
Peter Rotich authored
-
- Jan 23, 2015
-
-
Jared Hancock authored
Also try harder to send a relevant In-Reply-To and References header back to the client with the email message.
-
Peter Rotich authored
This pull request adds a cleanup util for bogus and invalid charsets, mostly added by a nameless company out of Redmond, WA.
-
- Jan 14, 2015
-
-
Jared Hancock authored
This patch sends updated session cookies to the browser when the session is refreshed on the server. This allows the session cookie to expire on the browser at the same time the session timeout occurs at the server. In the event the session timeout is configured in osTicket not to expire, the cookie will expire after seven days on the client browser, and will expire in PHP when it is garbage collected sometime after 86400 seconds after the time last refresh time. Using this method, the session will never expire if the session timeout in osTicket is configured to 0, and the session is refreshed at least daily.
-
- Jan 13, 2015
-
-
Chefkeks authored
Fixes https://github.com/osTicket/osTicket-1.8/issues/1673
-
Jared Hancock authored
Misc::randCode does not generate significantly random data for Windows platforms with a local database. This stems from the random seed using the milliseconds from the current time of day and the database connection time, in microseconds. Because Windows has especially poor sub-second time resolution via the microtime() function, the seed does not have many variations. This patch addresses the issue by using the included Crypto::random() function as a source of random data rather than the mt_rand() function, as it uses native cryptographic random data generators if possible to generate the data, and uses microtime() as a fallback if no other source of random data is available on the platform.
-
- Jan 12, 2015
-
-
Jared Hancock authored
-
- Jan 09, 2015
-
-
Jared Hancock authored
This fixes a slight issue where the team members would never be included on the new message alert. Now, the system will send to either the assigned staff member, if any, or the members of the assigned team, again, if any.
-
Jared Hancock authored
-
Jared Hancock authored
-
Jared Hancock authored
Files upgraded from versions of osTicket prior to 1.9.1 did not have a `signature` field in the database. For caching purposes in generating the Etag HTTP header, the getSignature function cascades through the getKey method. This may be inconsistent in the signed URL creation. This patch adds a cascade flag to the getSignature method so the cascading will not happen unless specifically requested.
-
- Jan 07, 2015
-
-
Jared Hancock authored
-
- Jan 06, 2015
-
-
Jared Hancock authored
Don't add to endTime if not already set
-
Jared Hancock authored
-
Jared Hancock authored
-
Jared Hancock authored
-
- Jan 05, 2015
-
-
Jared Hancock authored
-
- Jan 02, 2015
-
-
Peter Rotich authored
Show the form title on edit instead of "Custom Form" - which can be misleading since the user might be editing a built-in form.
-
- Dec 31, 2014
-
-
Jared Hancock authored
If a new ticket is assigned to both an individual agent and a team, do not send the email alert to the team lead or the team members.
-
Jared Hancock authored
-
Jared Hancock authored
This patch changes the ticket filter system so that it can be recursively looped to allow for banning and matching on the reply-to address and name.
-
Jared Hancock authored
This script adds a single download script, 'file.php', which provides access to files of all types to all users. It uses a HMAC signature system with an expires time, which allows signed URLs to be sent to external users. This also fixes an issue with the Http::cacheable() method, where the last-modified and Etag headers were not properly compared, which resulted in permanent cache misses by the client.
-
- Dec 30, 2014
-
-
Peter Rotich authored
Use http::build_query instead of inline urlencode
-
Jared Hancock authored
* Fix incorrect mapping to user email address * Fix early rejecting of tickets — even if a filter earlier in the matching filter list had "stop on match" set * Fix ::stopOnMatch referring to incorrect db field The new logic abandons the early rejection logic in ticket create. Instead, the normal validation is completed as usual. Thereafter, the filter is initialized and applied to the ticket. Upon rejection, a RejectedException is thrown by the ::apply() method of the TicketFilter. The Ticket::create() method will handle the exception and reject the ticket.
-
- Dec 28, 2014
-
-
Peter Rotich authored
Stop trampolining links via l.php. It was necessary before in order to avoid the potential of leaking ticket number & email. The authentication mechanism in place now redirects on successful login.
-
Peter Rotich authored
-
Peter Rotich authored
This is necessary so we can sanitize/encode inputs.
-
- Dec 26, 2014
-
-
Jared Hancock authored
This patch changes the default formatting for text bodies used in emails, ticket thread, and canned response quoting so that white-space in text bodies is properly preserved. Previously, the text was treated as raw HTML and was not properly escaped, nor was the original whitespace preserved.
-