Summary: Use in MailCommands and HTTP Parameters
Test Plan: Tested MailCommands in Paste, HTTP Parameters in Paste, Legalpad, Diviner. Mobile and Desktop breakpoints.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Differential Revision: https://secure.phabricator.com/D14397
Summary:
Ref T5873. Ref T9132. This is really rough and feels pretty flimsy at the edges (missing validation, generality, modularity, clean error handling, etc) but gets us most of the way toward generating plausible "whatever.edit" Conduit API methods from EditEngines.
These methods are full-power methods which can do everything the edit form can, automatically support the same range of operations, and update when new fields are added.
Test Plan:
- Used new `paste.edit` to create a new Paste.
- Used new `paste.edit` to update an existing paste.
- Applied a variety of different transactions.
- Hit a reasonable set of errors.
{F941144}
{F941145}
{F941146}
{F941147}
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T5873, T9132
Differential Revision: https://secure.phabricator.com/D14393
Summary:
Ref T9132. Although forms do generally support prefilling right now, you have to guess how to do it.
Provide an explicit action showing you which values are supported and how to prefill them. This is generated automatically when an application switches to ApplicationEditor.
Test Plan:
{F939804}
{F939805}
{F939806}
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T9132
Differential Revision: https://secure.phabricator.com/D14392
Summary:
Ref T9132. Ref T4768. This is a rough v0 of ApplicationEditor, which replaces the edit workflow in Paste.
This mostly looks and works like ApplicationSearch, and is heavily modeled on it.
Roughly, we define a set of editable fields and the ApplicationEditor stuff builds everything else.
This has no functional changes, except:
- I removed "Fork Paste" since I don't think it's particularly useful now that pastes are editable. We could restore it if users miss it.
- Subscribers are now editable.
- Form field order is a little goofy (this will be fixed in a future diff).
- Subscribers and projects are now race-resistant.
The race-resistance works like this: instead of submitting just the new value ("subscribers=apple, dog") and doing a set operation ("set subscribers = apple, dog"), we submit the old and new values ("original=apple" + "new=apple, dog") then apply the user's changes as an add + remove ("add=dog", "remove=<none>"). This means that two users who do "Edit Paste" at around the same time and each add or remove a couple of subscribers won't overwrite each other, unless they actually add or remove the exact same subscribers (in which case their edits legitimately conflict). Previously, the last user to save would win, and whatever was in their field would overwrite the prior state, potentially losing the first user's edits.
Test Plan:
- Created pastes.
- Created pastes via API.
- Edited pastes.
- Edited every field.
- Opened a paste in two windows and did project/subscriber edits in each, saved in arbitrary order, had edits respected.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T4768, T9132
Differential Revision: https://secure.phabricator.com/D14390
Summary: Ref T9352. See D13635. Build targets can have variables already, but let builds have them too. This mostly enables future use cases (sub-builds, more sophisticated build triggers).
Test Plan: With a custom Herald rule + action like the one in T9352, updated a revision and saw it generate multiple builds with varying parameters.
Reviewers: chad, hach-que
Reviewed By: hach-que
Maniphest Tasks: T9352
Differential Revision: https://secure.phabricator.com/D14222
Summary: Ref T9272. This doesn't fix anything, just a little cleanup while I was looking at it.
Test Plan: Clicked "Show Details" on a couple description changes, got the same effect for less code.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T9272
Differential Revision: https://secure.phabricator.com/D14168
Summary: Fixes T9369.
Test Plan:
- Sent a mail with Mail.app to `bugs@local.phacility.com`.
- Used "View Raw Mail", copy-pasted it into `mail.txt` on disk.
- Ran `cat mail.txt | ./scripts/mail/manage_mail.php --process-duplicates`.
- Saw task get created and me get added as CC.
- Changed "To" to include another user, ran command again, saw task get created and other user get added as CC.
Reviewers: chad
Reviewed By: chad
Subscribers: Korvin
Maniphest Tasks: T9369
Differential Revision: https://secure.phabricator.com/D14086
Summary:
Ref T8672. Ref T9187. Root issue in at least one case is:
- User makes a commit including a file with some non-UTF8 text (say, a Japanese file full of Shift-JIS).
- We pass the file to the TransactionEditor so it can inline or attach the patch if the server is configured for these things.
- When inlining patches, we convert them to UTF8 before inlining. We must do this since the rest of the mail is UTF8.
- When attaching patches, we send them in the original encoding (as file attachments). This is correct, and means we need to give the worker the raw patch in whatever encoding it was originally in: we can't just convert it to utf8 earlier, or we'd attach the wrong patch in some cases.
- TransactionEditor does its thing (e.g., creates the commit), then gets ready to send mail about whatever it did.
- The publishing work now happens in the daemon queue, so we prepare to queue a PublishWorker and pass it the patch (with some other data).
- When we queue workers, we serialize the state data with JSON.
So far, so good. But this is where things go wrong:
- JSON can't encode binary data, and can't encode Shift-JIS. The encoding silently fails and we ignore it.
Then we get to the worker, and things go wrong-er:
- Since the data is bad, we fatal. This isn't a permanent failure, so we continue retrying the task indefinitely.
This applies several fixes:
# When queueing tasks, fail loudly when JSON encoding fails.
# In the worker, fail permanently when data can't be decoded.
# Allow Editors to specify that some of their data is binary and needs special handling.
This is fairly messy, but some simpler alternatives don't seem like good ways forward:
- We can't convert to UTF8 earlier, because we need the original raw patch when adding it as an attachment.
- We could encode //only// this field, but I suspect some other fields will also need attention, so that adding a mechanism will be worthwhile. In particular, I suspect filenames //may// be causing a similar problem in some cases.
- We could convert task data to always use a serialize()-based binary safe encoding, but this is a larger change and I think it's correct that things are UTF8 by default, even if it makes a bit of a mess. I'd rather have an explicit mess like this than a lot of binary data floating around.
The change to make `LiskDAO` will almost certainly catch some other problems too, so I'm going to hold this until after `stable` is cut. These problems were existing problems (i.e., the code was previously breaking or destroying data) so it's definitely correct to catch them, but this will make the problems much more obvious/urgent than they previously were.
Test Plan:
- Created a commit with a bunch of Shift-JIS stuff in a file.
- Tried to import it.
Prior to patch:
- Broken PublishWorker with distant, irrelevant error message.
With patch partially applied (only new error checking):
- Explicit, local error message about bad key in serialized data.
With patch fully applied:
- Import went fine and mail generated.
Reviewers: chad
Reviewed By: chad
Subscribers: devurandom, nevogd
Maniphest Tasks: T8672, T9187
Differential Revision: https://secure.phabricator.com/D13939
Summary: Ref T8700, I don't believe we need to be specific here about the object, since it displays on the object.
Test Plan: Change policy a few times on a task, see new translation
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Maniphest Tasks: T8700
Differential Revision: https://secure.phabricator.com/D13913
Summary: Ref T5791. This is still very basic (no global actions, no support for matching headers/bodies/recipients/etc) but gets the core in.
Test Plan:
{F715209}
{F715211}
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T5791
Differential Revision: https://secure.phabricator.com/D13897
Summary: Use `PhutilClassMaQuery` instead of `PhutilSymbolLoader`, mostly for consistency. Depends on D13588.
Test Plan: Poked around a bunch of pages.
Reviewers: #blessed_reviewers, epriestley
Reviewed By: #blessed_reviewers, epriestley
Subscribers: epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D13589
Summary: Adds additional icon states for subscriber transactions. Also updated "eraser" to "trash" (man that icon is bad).
Test Plan: add a subscriber, remove a subscriber.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Differential Revision: https://secure.phabricator.com/D13800
Summary: Ref T8099, Cleans up UI issues, adds `appendList` and renders lists and paragraphs with Remarkup UI.
Test Plan: Test Policy Dialogs, other various dialogs.
Reviewers: btrahan, epriestley
Reviewed By: epriestley
Subscribers: epriestley, Korvin
Maniphest Tasks: T8099
Differential Revision: https://secure.phabricator.com/D13463
Summary: Fixes T8703. The URI handling here was a little unusual.
Test Plan:
- Edited and deleted comments in several applications, including Macro.
- As an admin, deleted others' comments.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8703
Differential Revision: https://secure.phabricator.com/D13469
Summary:
Ref T5791. This diff adds a "sensitive" flag to `PhabricatorMetaMTAMail`, defaults it to true in the constructor, and then sets it to false in teh application transaction editor. Assumption here is that sensitive emails are basically all the emails that don't flow through the application transaction editor.
This diff also gets a basic "mail view" page up and going.
This diff also fixes a bug writing recipient edges; the actor was being included.
This bug also fixes a querying bug; we shouldn't do the automagic join of $viewer is recipient or $viewer is actor if folks are querying for recipients or actors already. The bug manifested itself as having the "inbox" be inbox + outbox.
Test Plan: viewd list of messages. viewed message detail.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: epriestley, Korvin
Maniphest Tasks: T5791
Differential Revision: https://secure.phabricator.com/D13406
Summary:
Fixes T4139. Adds a "Desktop Notifications" panel to settings. For now, we start with "Send Desktop Notifications Too" functionality. We can try to be fancy later and only send desktop notifications if the web app doesn't have focus, etc.
Test Plan:
Made some comments as a test user on a task and got purdy desktop notifications using Chrome. Then did it again with Firefox.
Played around with permissions form with Chrome and got helpful information about what was up. Played around with Firefox and got similar results, except canceling the dialogue didn't invoke my handler code somehow. Oh Firefox!
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: rbalik, tycho.tatitscheff, joshuaspence, epriestley, Korvin
Maniphest Tasks: T4139
Differential Revision: https://secure.phabricator.com/D13219
Summary:
Ref T8574. Currently, failures during mail body construction, feed publishing, or search indexing could cause us to retry the publishing task and potentially send duplicate mail.
Instead, build (but do not send) the mail first, then send all the mail at the very end.
This isn't completley perfect, but should make it enormously harder for duplicate mail to be generated.
Test Plan: Sent some mail, ran the daemons, saw it show up normally in the outbound queue.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8574
Differential Revision: https://secure.phabricator.com/D13320
Summary: All classes should extend from some other class. See D13275 for some explanation.
Test Plan: `arc unit`
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D13283
Summary:
Ref T5681. Getting this to work correctly is a bit tricky, mostly because of the policy checks we do prior to applying an edit.
I think I came up with a mostly-reasonable approach, although it's a little bit gross. It uses `spl_object_hash()` so it shouldn't be able to do anything bad/dangerous (the hints are strictly bound to the hinted object, which is a clone that we destroy moments later).
Test Plan:
- Added + ran a unit test.
- Created a task with a "Subscribers" policy with me as a subscriber (without the hint stuff, this isn't possible: since you aren't a subscriber *yet*, you get a "you won't be able to see it" error).
- Unsubscribed from a task with a "Subscribers" policy, was immediately unable to see it.
- Created a task with a "subscribers" policy and a project subscriber with/without me as a member (error / success, respectively).
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T5681
Differential Revision: https://secure.phabricator.com/D13259
Summary:
Ref T5681. Ref T8488. This allows policy rules to provide "Object Policies", which are similar to the global/basic policies:
- They show up directly in the dropdown (you don't have to create a custom rule).
- They don't need to create or load anything in the database.
To implement one, you just add a couple methods on an existing PolicyRule that let Phabricator know it can work as an object policy rule.
{F494764}
These rules only show up where they make sense. For example, the "Task Author" rule is only available in Maniphest, and in "Default View Policy" / "Default Edit Policy" of the Application config.
This should make T8488 easier by letting us set the default policies to "Members of Thread", without having to create a dedicated custom policy for every thread.
Test Plan:
- Set tasks to "Task Author" policy.
- Tried to view them as other users.
- Viewed transaction change strings.
- Viewed policy errors.
- Set them as default policies.
- Verified they don't leak into other policy controls.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T5681, T8488
Differential Revision: https://secure.phabricator.com/D13257
Summary:
Ref T8498. Allow ApplicationEmail addresses to be put into spaces:
- You can only see and send to addresses in Spaces you have access to.
- Objects are created into the same space their address is associated with.
Test Plan:
- Used `bin/mail receive-test` to send mail to various `xyz-bugs@...` addresses.
- Saw objects created in the proper space.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8498
Differential Revision: https://secure.phabricator.com/D13247
Summary:
Ref T8498. I want to add Spaces to these, and the logic for getting Spaces right is a bit tricky, so swap these to ApplicationTransactions.
One new piece of tech: made it easier for Editors to raise DuplicateKeyException as a normal ValidationException, so callers don't have to handle this case specially.
One behavioral change: we no longer require these addresses to be at the `auth.email-domains` domains -- I think this wasn't quite right in the general case. It's OK to require users to have `@mycompany.com` addresses but add `@phabricator.mycompany-infrastructure.com` addresses here if you want.
Test Plan:
- Tried to create a duplicate email.
- Tried to create an empty email.
- Tried to create an invalid email.
- Created a new email.
- Deleted an email.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8498
Differential Revision: https://secure.phabricator.com/D13246
Summary:
Ref T8377. This adds a standard disable/enable feature to Spaces, with a couple of twists:
- You can't create new stuff in an archived space, and you can't move stuff into an archived space.
- We don't show results from an archived space by default in ApplicationSearch queries. You can still find these objects if you explicitly search for "Spaces: <the archived space>".
So this is a "put it in a box in the attic" sort of operation, but that seems fairly nice/reasonable.
Test Plan:
- Archived and activated spaces.
- Used ApplicationSearch, which omitted archived objects by default but allowed searches for them, specifically, to succeed.
- Tried to create objects into an archived space (this is not allowed).
- Edited objects in an archived space (this is OK).
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8377
Differential Revision: https://secure.phabricator.com/D13238
Summary: Fixes T8464. We could lose the additional users from "Send an email..." rules //if// Herald did not apply any other transactions to the task.
Test Plan:
- Destroyed all Herald rules.
- Created a single "Send an email to..." rule.
- Created a task.
- Saw target get an email.
Reviewers: btrahan, chad
Reviewed By: chad
Subscribers: epriestley
Maniphest Tasks: T8464
Differential Revision: https://secure.phabricator.com/D13245
Summary:
Ref T8449. Try out some more subtle behaviors:
- Make the "Space" control part of the policy control, so the UI shows "Visible To: [Space][Policy]". I think this helps make the role of spaces more clear. It also makes them easier to implement.
- Don't show the default space in headers: instead, show nothing.
- If the user has access to only one space, pretend spaces don't exist (no edit controls, no header stuff).
This might be confusing, but I think most of the time it will all align fairly well with user expectation.
Test Plan:
- Viewed a list of pastes (saw Space with non-default space, no space with default space, no space with user in only one space).
- Viewed a paste (saw Space with non-default space, saw no space with default space, saw no space with user in only one space).
- Edited spaces on objects (control as privileged user, no control as locked user).
- Created a new paste in a space (got space select as privileged user, no select as locked user).
Reviewers: chad, btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8449
Differential Revision: https://secure.phabricator.com/D13229
Summary:
Fixes T8483. I did this incorrectly in D13159, by doing it correctly first and then editing it carelessly. For most transaction types, it didn't matter, but did for inline state.
Also, clean up any bad inline state transactions.
Test Plan:
- Ran migration, bad transactions vanished.
- Marked some inline comments as done.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8483
Differential Revision: https://secure.phabricator.com/D13226
Summary: Now that Users implement PhabricatorApplicationTransactionInterface, we try to write an inverse edge. At least for now, we should retain the old behavior instead.
Test Plan:
- Unit tests which cover this stuff pass again.
- Grepped for other `instanceof PhabricatorApplicationTransactionInterface`, the all seemed either benign or irrelevant.
Reviewers: joshuaspence, btrahan
Reviewed By: btrahan
Subscribers: epriestley
Differential Revision: https://secure.phabricator.com/D13215
Summary:
Fixes T8464. We could incorrectly use a cached value when computing CC's.
Just load a fresh value. There are no other callers that would benefit from this cache, so it's more complicated to reload it correctly prior to publishing than to just skip it.
Also make the PHID headers unique.
Test Plan:
- Verified that users received mail about the transactions which caused them to be added to an object.
- Veirfied that headers no longer have redundant values.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8464
Differential Revision: https://secure.phabricator.com/D13206
Summary:
Ref T8424. This adds crude integration with Paste's edit/view workflows: you can change the space a Paste appears in, see transactions, and get a policy callout.
Lots of rough edges and non-obviousness but it pretty much works.
Test Plan:
- Created and updated Pastes.
- Moved them between spaces, saw policy effects.
- Read transactions.
- Looked at feed.
- Faked query to return no spaces, saw control and other stuff vanish.
- Faked query to return no spaces, created pastes.
- Tried to submit bad values and got errors.
Reviewers: chad, btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8424
Differential Revision: https://secure.phabricator.com/D13159
Summary: Ref T6367.
Test Plan:
- Added and executed unit tests.
- Sent mail to A (en_US) and B (en_A*).
- Got one mail in English and one mail in ENGLISH.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T6367
Differential Revision: https://secure.phabricator.com/D13142
Summary:
Ref T6367. Removes `multiplexMail()`!
We can't pass a single body into a function which splits it anymore: we need to split recipients first, then build bodies for each recipient list. This lets us build separate bodies for each recipient's individual translation/access levels.
The new logic does this:
- First, split recipients into groups called "targets".
- Each target corresponds to one actual mail we're going to build.
- Each target has a viewer (whose translation / access levels will be used to generate the mail).
- Each target has a to/cc list (the users who we'll ultimately send the mail to).
- For each target, build a custom mail body based on the viewer's access levels and settings (language prefs not actually implemented).
- Then, deliver the mail.
Test Plan:
- Read new config help.
Then did a bunch of testing, primarily with `bin/mail list-outbound` and `bin/mail show-outbound` (to review generated mail), `bin/phd debug taskmaster` (to run daemons freely) and `bin/worker execute --id <id>` (to repeatedly test a specific piece of code after identifying an issue).
With `one-mail-per-recipient` on (default):
- Sent mail to multiple users.
- Verified mail showed up in `mail list-outbound`.
- Examined mail with `mail show-outbound`.
- Added a project that a subscriber could not see.
- Verified it was not present in `X-Phabricator-Projects`.
- Verified it was rendered as "Restricted Project" for the non-permissioned viewer.
- Added a subscriber, then changed the object policy so they could not see it and sent mail.
- Verified I received mail but the other user did not.
- Enabled public replies and verified mail generated with public addresses.
- Disabld public replies and verified mail generated with private addresses.
With `one-mail-per-recipient` off:
- Verified that one mail is sent to all recipients.
- Verified users who can not see the object are still filtered.
- Verified that partially-visible projects are completely visible in the mail (this violates policies, as documented, as the best available compromise).
- Enabled public replies and verified the mail generated with "Reply To".
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: carlsverre, epriestley
Maniphest Tasks: T6367
Differential Revision: https://secure.phabricator.com/D13131
Summary:
Ref T6367. Do all mail, feed, notification and search stuff from the daemons, in all editors.
There are four relatively-stateful editors (Audit, Differential, Phriction, PhortuneCart) which needed special care to move state into the daemons properly.
Beyond that, I moved mailTo/mailCC/feedRelated/feedNotify to be computed before we enter the worker:
- This is simpler, since a lot of editors rely on being able to call `$object->getReviewers()` or similar to compute them.
- This is more correct, since we want to freeze the lists at this moment in time.
Finally, I renamed `loadEdges` to `willPublish` and made it a slightly more general hook.
---
This is a bit fragile and I'm not //thrilled// about it.
It would probably be cleaner to have separate Editor and Publisher classes (something like @fabe's D11329 did). However, I think that's quite a lot of work, and I'd like to see stronger motivation for it (either in this actually being more fragile than I think, or there being other things we get out of it). Overall, I'm comfortable with this change, just definitely not a big fan of the "save" + "load" pattern since I think it's really fragile, nonobvious, hard to debug/predict, etc.
Test Plan:
Directly updated editors:
- Created a new Phriction page, saw "Document Content".
- Edited a Phriction page, saw "Document Diff".
- Edited a revision, got normal looking mail.
- Faked in `changedPriorToCommitURI` and verified it survived the state boundary.
- Sent Audit mail.
- Sent invoice mail.
Indirect editors - for these, I just made a change and made sure the mail generated:
- Updated a paste.
- Updated an event.
- Updated a thread.
- Updated a task.
- Updated a mock.
- Updated a question.
- Updated a project.
- Updated a file.
- Updated an initiative.
- Updated a Legalpad document.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley, fabe
Maniphest Tasks: T6367
Differential Revision: https://secure.phabricator.com/D13115
Summary:
Ref T6367. This is similar to D11329, but not quite as ambitious.
Allow Editors to implement `supportsWorkers()` and move their publishing work into a daemon. So far, only Paste supports this.
Most of the complexity here is saving and restoring state across the barrier between the web process and the worker process, but I think this is ~90% of it and then we'll pick up a couple of random things in applications.
I'm primarily trying to keep this as gradual as possible.
Test Plan:
- Published transactions with and without daemon support.
- Looked at mail, feed.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: fabe, epriestley
Maniphest Tasks: T6367
Differential Revision: https://secure.phabricator.com/D13107
Summary: Fixes T8329. I was able to figure out a reasonable way to have the full conpherence default to the email settings panel. I think this is cleaner than making things a dialogue as I rambled about in the description for T8329.
Test Plan: using /bin/mail to verify correct email links were generated for conpherence notifications and maniphest (general) notifications.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin, epriestley
Maniphest Tasks: T8329
Differential Revision: https://secure.phabricator.com/D13058
Summary: Fixes T6403. Remaining built-ins were already built-in effectively so this is a small re-factor plus some docs. I probably wouldn't have written anything if not for the TODO so please feel free to tell me to write something else or what have you.
Test Plan: NA since this didn't actually change anything.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin, epriestley
Maniphest Tasks: T6403
Differential Revision: https://secure.phabricator.com/D12937
Summary: Ref XXXXX. I broke things a bit in XXXXX in that if the TYPE_EDGE had an inverse transaction, we weren't correctly "doing nothing" and instead were falling back to our old every editor has to implement a no-op ways... Fix things by putting the TYPE_EDGE code in the handle external builtin function like it belongs.
Test Plan: made a comment on a task referencng a commit successfully
Reviewers: epriestley
Subscribers: Korvin, epriestley
Maniphest Tasks: T6403
Differential Revision: https://secure.phabricator.com/D12939
