3 ms·
It should be patched since Dec 2019: https://twitter.com/jensvoid/status/1295633279727673344 https://twitter.com/jensvoid/status/1295633279727673344
by miken123 6y ago
It should be patched since Dec 2019: https://twitter.com/jensvoid/status/1295633279727673344 https://twitter.com/jensvoid/status/1295633279727673344
- avian 6y agoI found a discussion from March 2009 which suggests this issue was known back then [1]. Maybe the same feature was re-introduced later in some other part of the code? The link to source repo in the post is dead, but the file it references (nsSmtpUrl.cpp) still contains this comment on line 88 in current version: /* DO NOT support attachment= in mailto urls. This poses a security fire hole!!! [1] http://forums.mozillazine.org/viewtopic.php?f=30&t=1172555 http://forums.mozillazine.org/viewtopic.php?f=30&t=1172555
- m4tthumphrey 6y agoThat interesting as the OP is using `attachment` as the param when it is/was actually `attach`.
- TheSoftwareGuy 6y agoLooking at the linked src code, there’s lots of mixed tabs/spaces for indent. shudders
- sukilot 6y agoA comment in source code. Not the discerning professional's preferred tool for guaranteeing security-critical behaviors to prevent "fire holes".
- setzer22 6y agoWhat should they do? Static analysis would be nice, but it's probably too complex to adopt in that kind of codebase. In the end, the only viable option is to forbid this behaviour, and document a rationale so nobody will enable it back again. Perhaps they could make this warning bigger, but other than that...
- jefftk 6y agoThey could have added a regression test, no? Give it an example with ?attach= and verify that no attachment is added.