Summary:
See <https://github.com/facebook/phabricator/issues/541>.
- If a provider returns the email `""` or `"0"`, we currently don't let the user edit it and thus don't let them register.
- If a provider returns an invalid email like `"!!!"` (permitted by GitHub, e.g.), we show them a nonsense error message.
Instead:
- Pretend we didn't get an address if we get an invalid address.
- Test the address strictly against `null`.
Test Plan: Registered on Phabricator with my GitHub email set to `""` (empty string) and `"!!!"` (bang bang bang).
Reviewers: btrahan, chad
Reviewed By: chad
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8528
Summary: This way the type of story can be inferred.
Test Plan: requested feed.query with `view=>'text'`
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: aran, epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D8521
Summary: Fixes T4600. If there's also a revision, the variable "$message" gets overwritten. groan~
Test Plan: Pushed a commit with "Fixes T123" and a revision, saw it parse on the first try.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: chrisbolt, aran, epriestley
Maniphest Tasks: T4600
Differential Revision: https://secure.phabricator.com/D8519
Summary:
This revision adds a 'method' field to the HTTP request harbormaster build step. This allows the user to specify GET, POST, DELETE, and PUT (limited by the underlying wrapper phabricator uses for HTTP requests). I'm not sure how much sense PUT makes, but oh well.
Existing plans shouldn't break, as if this field is an empty string, we default to POST, which is the old behavior.
Fixes T4604
Test Plan: 1) Verified that the empty string does, in fact, issue a POST request. Changed the method to be GET and observed that the problem described in T4604 is resolved.
Reviewers: #blessed_reviewers, epriestley
Reviewed By: #blessed_reviewers, epriestley
Subscribers: aran, epriestley
Maniphest Tasks: T4604
Differential Revision: https://secure.phabricator.com/D8520
Summary:
Fixes T4362. If you have a default edit + view policy of "no one" things get weird when you try to create a task - basically its impossible.
Ergo, re-jigger how we do policy checks just a bit.
- if its a new object, don't bother with the "can the $actor edit this thing by virtue of having can see / can edit priveleges?" That makes no sense on create.
- add a hook so when doing the "will $actor still be able to edit this thing after all the edits" checks the object can be updated to its ultimate state. This matters for Maniphest as being the owner lets you do all sorts of stuff.
Test Plan:
- made a task with no one policy and assigned to no one - exception
- made a task with no one policy and assigned to me - success
- made a comment on the task - success
- reassigned the task to another user - exception
- reassigned the task to another user and updated policies to "users" - success
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: aran, epriestley, Korvin
Maniphest Tasks: T4362
Differential Revision: https://secure.phabricator.com/D8508
Summary:
Ref T4593. Via HackerOne. An attacker can use the anchor reattachment, combined with the Facebook token workflow, combined with redirection on OAuth errors to capture access tokens. The attack works roughly like this:
- Create an OAuth application on Phabricator.
- Set the domain to `evil.com`.
- Grab the OAuth URI for it (something like `https://phabricator.com/oauthserver/auth/?redirect_uri=http://evil.com&...`).
- Add an invalid `scope` parameter (`scope=xyz`).
- Use //that// URI to build a Facebook OAuth URI (something like `https://facebook.com/oauth/?redirect_uri=http://phabricator.com/...&response_type=token`).
- After the user authorizes the application on Facebook (or instantly if they've already authorized it), they're redirected to the OAuth server, which processes the request. Since this is the 'token' workflow, it has auth information in the URL anchor/fragment.
- The OAuth server notices the `scope` error and 302's to the attacker's domain, preserving the anchor in most browsers through anchor reattachment.
- The attacker reads the anchor in JS and can do client workflow stuff.
To fix this, I've made several general changes/modernizations:
- Add a new application and make it beta. This is mostly cleanup, but also turns the server off for typical installs (it's not generally useful quite yet).
- Add a "Console" page to make it easier to navigate.
- Modernize some of the UI, since I was touching most of it anyways.
Then I've made specific security-focused changes:
- In the web-based OAuth workflow, send back a human-readable page when errors occur. I //think// this is universally correct. Previously, humans would get a blob of JSON if they entered an invalid URI, etc. This type of response is correct for the companion endpoint ("ServerTokenController") since it's called by programs, but I believe not correct for this endpoint ("AuthController") since it's used by humans. Most of this is general cleanup (give humans human-readable errors instead of JSON blobs).
- Never 302 off this endpoint automatically. Previously, a small set of errors (notably, bad `scope`) would cause a 302 with 'error'. This exposes us to anchor reattachment, and isn't generally helpful to anyone, since the requesting application did something wrong and even if it's prepared to handle the error, it can't really do anything better than we can.
- The only time we'll 'error' back now from this workflow is if a user explicitly cancels the workflow. This isn't a 302, but a normal link (the cancel button), so the anchor is lost.
- Even if the application is already approved, don't blindly 302. Instead, show the user a confirmation dialog with a 'continue' link. This is perhaps slightly less user-friendly than the straight redirect, but I think it's pretty reasonable in general, and it gives us a lot of protection against these classes of attack. This redirect is then through a link, not a 302, so the anchor is again detached.
-
Test Plan: I attempted to hit everything I touched. See screenshots.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4593
Differential Revision: https://secure.phabricator.com/D8517
Summary:
Via HackerOne. In regular expressions, "$" matches "end of input, or before terminating newline". This means that the expression `/^A$/` matches two strings: `"A"`, and `"A\n"`.
When we care about this, use `\z` instead, which matches "end of input" only.
This allowed registration of `"username\n"` and similar.
Test Plan:
- Grepped codebase for all calls to `preg_match()` / `preg_match_all()`.
- Fixed the ones where this seemed like it could have an impact.
- Added and executed unit tests.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8516
Summary:
Ref T4587.
- Add an option to generate a keypair.
- Add an option to view the public keys for existing keypairs.
Test Plan:
- Generated keypairs.
- Viewed public keys.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4587
Differential Revision: https://secure.phabricator.com/D8515
Summary: Ref T4587. Add an option to automatically generate a keypair, associate the public key, and save the private key.
Test Plan: Generated some keypairs. Hit error conditions, etc.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4587
Differential Revision: https://secure.phabricator.com/D8513
Summary:
Currently, disabling Herald only disables feed, notifications and email. Historically, audits didn't really create external effects so it made sense for Herald to only partially disable itself.
With the advent of Harbormaster/Build Plans, it makes more sense for Herald to just stop doing anything. When this option is disabled, stop all audit/build/publish/feed/email actions for the repository.
Test Plan: Ran `scripts/repository/reparse.php --herald`, etc.
Reviewers: dctrwatson, btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8509
Summary: In the Message parser, we read this field and expect to get an array of PHIDs out of it. Currently, we get a string. Instead, get an array of PHIDs.
Test Plan: Wrote a message like "Fixes Tnnn" with "Reviewed by: duck", and saw no more parse error during message parsing.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8510
Summary: Similar to D8491, some of my new exceptions are a bit too aggressive. See IRC. This one's hitting an edit workflow with 'revisionID' onboard somehow.
Test Plan: Not entirely sure how to hit this, but it won't throw anymore.
Reviewers: btrahan, dctrwatson, zeeg
Reviewed By: zeeg
Subscribers: zeeg, aran, epriestley
Differential Revision: https://secure.phabricator.com/D8514
Summary: "Users who an edit" to "Users who can edit"
Test Plan: Verified that typo is gone after the change
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: aran, epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D8511
Summary: For some actions (like accept) we need to load reviewer authority so we can figure out if the actor can act on behalf of project reviewers, etc.
Test Plan: Will make @dctrwatson do it.
Reviewers: dctrwatson, btrahan
Reviewed By: btrahan
Subscribers: chad, aran, dctrwatson, epriestley
Differential Revision: https://secure.phabricator.com/D8505
Summary:
Two issues:
- Herald is currently overwriting accepts and rejects with "blocking reviewer". Just stop it from doing that.
- When you update an accepted revision, we put it back in "needs review", then return it to "accepted", generating an extra transaction. Instead, don't.
Test Plan:
- Updated a revision with an accepting, herald-based blocking project reviewer. Reviewer was still accepting.
- Updated an accepted revision, didn't get an extra transaction.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8506
Summary:
Fixes T4594. Also, allow "exists" / "does not exist" to be run against author/committer. This allows construction of rules like:
- Committer identities must be authentic.
- Committer identities must be resolvable.
- Author identities must be resolvable.
Test Plan: Created some rules using these new rules and ran them.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4594
Differential Revision: https://secure.phabricator.com/D8507
Summary:
Ref T4585.
- Use modern UI kit.
- Make mobile-ish.
- Fix a couple minor things.
Test Plan:
{F127155}
{F127156}
Reviewers: btrahan, chad
Reviewed By: chad
Subscribers: aran, epriestley, chad
Maniphest Tasks: T4585
Differential Revision: https://secure.phabricator.com/D8504
Summary:
- Fixes T4588.
- See D8501.
- Adds a "Tags" field for Herald commit emails.
- Fixes a bug in `tagsquery` when filtering by commit name.
- Make `tagsquery` just return nothing instead of fataling against Mercurial/Subversion.
Test Plan: Used `bin/repository/reparse.php --herald` to exercise this code.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4588
Differential Revision: https://secure.phabricator.com/D8502
Summary:
Ref T4588. Request from @zeeg. Adds a "BRANCHES" field to commit emails, so the branches the commit appears on are shown.
I've implemented this with CustomField, but in a very light way.
Test Plan: Used `scripts/repository/reparse.php --herald` to generate mail, got a BRANCHES section where applicable.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley, zeeg
Maniphest Tasks: T4588
Differential Revision: https://secure.phabricator.com/D8501
Summary:
Ref T4592. These were added with the intent of not requiring builds on Windows, but then we got builds on Windows working and they seem to be straightforward. See T4592 for most recent discussion.
Remove these methods because they aren't really practical for anything and increase attack surface area by giving adversaries access to `xhpast`, and generally bloat up the Conduit API. To my knowledge, nothing has ever called them.
(If an install somehow relies on these, they can drop them into `src/extensions/` to expose them again.)
Test Plan: Viewed conduit.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4592
Differential Revision: https://secure.phabricator.com/D8500
Summary:
Ref T4593. There are a variety of clever attacks against OAuth which involve changing the redirect URI to some other URI on the same domain which exhibits unexpected behavior in response to an OAuth request. The best approach to dealing with this is for providers to lock to a specific path and refuse to redirect elsewhere, but not all providers do this.
We haven't had any specific issues related to this, but the anchor issue in T4593 was only a step away.
To mitigate this in general, we can reject the OAuth2 `'code'` parameter on //every// page by default, and then whitelist it on the tiny number of controllers which should be able to receive it.
This is very coarse, kind of overkill, and has some fallout (we can't use `'code'` as a normal parameter in the application), but I think it's relatively well-contained and seems reasonable. A better approach might be to whitelist parameters on every controller (i.e., have each controller specify the parameters it can receive), but that would be a ton of work and probably cause a lot of false positives for a long time.
Since we don't use `'code'` normally anywhere (as far as I can tell), the coarseness of this approach seems reasonable.
Test Plan:
- Logged in with OAuth.
- Hit any other page with `?code=...` in the URL, got an exception.
- Grepped for `'code'` and `"code"`, and examined each use to see if it was impacted.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4593
Differential Revision: https://secure.phabricator.com/D8499
Summary:
Fixes T4586, where a keystroke like "return" to submit a form could remove DOM nodes rendering a tooltip, leaving the user with a permanent phantom tooltip.
(@spicyj, if you come up with anything better than this feel free to send a followup, but I //think// this is a reasonable fix.)
A nice side effect of this is that it hides any tooltips obscuring a text input when you start typing.
Test Plan: Hovered over a tooltip, pressed a key, tip vanished.
Reviewers: spicyj, btrahan
Reviewed By: btrahan
Subscribers: aran, spicyj, epriestley
Maniphest Tasks: T4586
Differential Revision: https://secure.phabricator.com/D8497
Summary: This is required for browsers <= IE7
Test Plan: Inspected HTML for type.
Reviewers: epriestley
Subscribers: aran, epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D8503
Summary: Ref T2217. I'll hold this for a month or so, but once we're confident the migration didn't ruin anything we should nuke this old data -- it's just an insurance policy against discovering migration issues.
Test Plan: Will run in a month or so.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley, aran
Maniphest Tasks: T2217
Differential Revision: https://secure.phabricator.com/D7104
Summary:
This is a bit messy, but not tooo bad:
- In general, stop the author from being added as a reviewer.
- In the specific case that "self accept" is enabled, allow it. This is easier than trying to special case it.
- When commandeering, we make the author a reviewer and make the actor the author, but these happen after validation. At validation time, it looks like we're making the author a reviewer. Just special case this.
- Provide a slightly nicer message when trying to add yourself from `arc`. We hit the Transactions message anyway, but it's not formatted as cleanly.
- Don't try to add the author via Herald.
Test Plan:
- Edited a revision with author = reviewer, got stopped.
- Commandeered revision.
- Updated from `arc`.
- Updated in general.
- Fired a "add author as reviewer" Herald rule without things breaking.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8496
Summary: This got dropped in the ApplicationTransactions stuff.
Test Plan: Created a new revision.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8495
Summary:
Fixes two issues with Differential:
- New reviewers on initial diff were being created into a `null` state.
- The `"="` edge update was overwriting accepted/rejected statuses. This could maybe be more nuanced in the long run, but I've just made it update correctly for now.
Test Plan:
- Created and updated a revision, paying attention to reviewer statuses.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8494
Summary: Via HackerOne. This doesn't actually have any security impact as far as we can tell, but a researcher reported it since it seems suspicious. At a minimum, it could be confusing. Also improve some i18n stuff.
Test Plan: Hit all the error cases, then saved a valid custom domain.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8493
Summary: see title. Fixes T4549.
Test Plan: made a readme that had some headers and observed a nice ToC
Reviewers: chad, epriestley
Reviewed By: epriestley
Subscribers: aran, epriestley, Korvin
Maniphest Tasks: T4549
Differential Revision: https://secure.phabricator.com/D8490
Summary:
nothing too crazy here. try to be smart about some defaults (i.e. phame title is optional and can be derived from title; post as not a draft by default; etc). Fixes T3695.
also do a little re-factoring to centralizing initializing new posts and turning posts into dictionaries. also change blogs => posts in another conduit method so it makes sense and stuff.
Test Plan: made some posts via conduit. testing trying to specify blogger, phame title, and isDraft, all worked nicely
Reviewers: chad, epriestley
Reviewed By: epriestley
Subscribers: aran, epriestley, Korvin
Maniphest Tasks: T3695
Differential Revision: https://secure.phabricator.com/D8485
Summary: Hit this issue in D8485. I think reviewedByPHID changes should appear in application transactions.
Test Plan: Would like to deploy this and try updating D8485 again.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: aran, epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D8491
Summary: The JIRA field is currently always enabled. This isn't correct; it
should be disabled if there's no JIRA provider.
We also use the old set of reviewers to compute mail delivery. Instead, reload
the correct set of reviewers before moving on after finalizing transactions.
Summary: Useful in cases where there is an Arcanist Project but not a repository tracked by Phabricator for a particular revision.
Test Plan: Created a new rule to flag Differential revisions with a particular Arcanist project, verified that it applied as expected via the test console to revisions with the project specified and with a different project specified.
Reviewers: #blessed_reviewers, epriestley
Reviewed By: #blessed_reviewers, epriestley
Subscribers: aran, epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D8463
Summary:
It appears a change to the way the configuration was loaded into ArcanistRepositoryAPI in rARCa2285b2b broke the save_lint script.
This updates the DiffusionLintSaveRunner to use the configuration correctly, allowing the linter to run
Test Plan: cd /your/project; ../../../path/to/phabricator/scripts/repository/save_lint.php
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: aran, epriestley, Korvin
Differential Revision: https://secure.phabricator.com/D8487
Summary:
Ref T2222.
- Removes `DifferentialTasksAttacher`, which has had no callsites for a very long time.
- Moves `differential.getrevisioncomments` off `DifferentialCommentQuery`.
- Moves Releeph churn field off `DifferentialCommentQuery`.
- Removes dead code in `DifferentialRevisionViewController`.
- Removes `DifferentialException` (no references).
- Removes `DifferentialRevision->loadComments()` (no callsites).
- Removes `DifferentialRevision->loadReviewedBy()` (all callsites updated).
- Removes `DifferentialCommentQuery` (all callsites updated).
Test Plan: Mostly a lot of `grep`.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8476
Summary: Ref T2222. Makes the "lint/unit errors" warnings work again.
Test Plan: Viewed some revisions with and without these warnings.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8475
Summary: Ref T2222. The unit and lint fields still have one piece of functionality that I need to port, but everythign else is obsolete.
Test Plan: Lots of `grep`.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8474
Summary: Ref T2222. Straightforward, just breaks a needless dependency.
Test Plan: Pushed and parsed a commit with "Auditors" in it.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8473
Summary: Ref T2222. There's some magic here, just port it forward in a mostly-reasonable way. This could use some refinement eventually.
Test Plan: Pushed commits with "Fixes" and "Ref" language, used `reparse.php` to trigger the new code. Saw expected updates in Maniphest.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8471
Summary:
Ref T2222. This has some minor functionality regressions:
- The plain diff page no longer shows unit/test status. I want to give diffs separate custom fields for this.
- It was technically possible to shove more data on the list view, although this doensn't affect the default config.
Test Plan: Looked at list view, diff detail view. Grepped for changes.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8470
Summary: Ref T2222. These no longer have an effect, and are obsoleted by `differential.fields`.
Test Plan: `grep`
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8468
Summary:
Ref T2222. Brings the major mail features (affected files, patches) forward.
This drops some of the minor integrations which just show object state (like "Maniphest Tasks") since I think they're not very important. I'll put them back if users miss them.
Test Plan: Sent mail with inline/attached patches.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8459
