This does introduce a small security issue, in that a shell with
expired AFS tokens on the machine running the mirroring script will be
able to read the Humbug API key using /proc/pid/environ. I think this
is fine -- you can steal the API key from a running process using
ptrace anyway.
(imported from commit c6fdb798294fb32d640823b409f3e46274ca01f4)
After this commit, we're tracking two branches: "master" and "prod".
Pushed to "prod" deploy to production, while pushes to "master" deploy
to the staging server.
(imported from commit 6ce429a1d6f606fb6136341dc393d93fd1228a21)
This was causing us to log some requests twice, and might have more serious
consequences as well.
(imported from commit 0bb2d7207ee3e4e04679215a7f5ae637cd26aa19)
The zero-port case never actually worked, because addrport wasn't an optional
parameter in run_one. And multiple ports was implemented using the
multiprocessing library, which is just bad news. Since we have no need for
this, remove it before it can cause trouble.
(imported from commit 9d913924701f30d23ebe878b76c8f1f0da2800e2)
This seems to fix the page-scroll issue and is also a little less
ostentatious.
Fixes https://trac.humbughq.com/ticket/243
(imported from commit 5fdbcc62a251303d134aea46a0755984df125c63)
Because in the zephyr world, people often manually linewrap their
zephyrs, we need to relax the algorithm's assumption that every line
was wrapped to the same linewrapping threshhold. Otherwise messages
that look like this:
aaaaa aaaaa aaaaa aaaaaaaaaa aaaaa aaaaa aaaaa aaaaa
aaaaa aaaaa aaaaa aaaaaaaaaa aaaaa aaaaa aaaaa aaaaa aaaaa aaaaaa
aaaaa aaaaa aaaaa aaaaaaaaaa aaaaa aaaaa aaaaa aaaaa
might be treated as a paragraph break, even though that was unlikely
to have been the sender's intention.
(imported from commit e1144da2a3efc3241a8f15b0f19fea2ea6684254)
I think it might be nice to have it also specify some information
about the new review request, but I think we benefit a lot from
getting this minimal version out the door.
(imported from commit cb04c72ee125763cff2d83a6afad3f5319d8ebc6)
It's generally best to first try killing with SIGTERM, and if we fail
to kill the child, we'll end up trying to kill them with SIGKILL
anyway when the new zephyr_mirror process starts up,
(imported from commit cfee2dd5f809f6e38d90a09be82719a8660d8377)
The comment on keydown_handler says that these functions should
"return a new handler, or 'false' to decline to handle the event."
(imported from commit 8cd23ee69ef900fcb7c7c211fe6ad36f54f02ba9)
This reverts commit 429e055d3eca65af8bc0fe58481a7becf9ced66a.
There is some inconsistency between the names 'huddle' and 'personal' that is
breaking things.
(imported from commit 4c81853fca9d88d13ce8f23e2d6884c33cdc57d2)
This means you can click the label to focus the corresponding box.
Also, it's more consistent with our other HTML, and may have accessibility
benefits.
(imported from commit c40de1e4137eb8ce5c69775ebcae346af9ae4e32)
Keegan McAllister
Tim Abbott
Waseem Daher
Zev Benjamin
Jeff Arnold