Summary:
This is partly a good feature, and partly should reduce false positives on HackerOne reporting things vaguely related to this.
Allow a user to terminate login sessions from the settings panel.
Test Plan:
- Terminated a session.
- Terminated all sessions.
- Tried to terminate all sessions again.
- Logged in with two browsers, terminated the other browser's session, reloaded, got kicked out.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Differential Revision: https://secure.phabricator.com/D8556
Summary:
- Point them at the new Diviner.
- Make them a little less cumbersome to write.
Test Plan: Found almost all of these links in the UI and clicked them.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Differential Revision: https://secure.phabricator.com/D8553
Summary:
This is the other half of D8548. Specifically, the attack here was to set your own editor link to `javascript\n:...` and then you could XSS yourself. This isn't a hugely damaging attack, but we can be more certain by adding a whitelist here.
We already whitelist linkable protocols in remarkup (`uri.allowed-protocols`) in general.
Test Plan:
Tried to set and use valid/invalid editor URIs.
{F130883}
{F130884}
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Differential Revision: https://secure.phabricator.com/D8551
Summary:
Fixes T4619. Currently, even if a viewer can't see Maniphest, they'll still see empty panels on the home page. These panels will always be empty so there's no real policy violation, but it's confusing.
Longer term, dashboards should fix this.
Test Plan: Viewed home page with a user with and without permissions on the apps.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T4619
Differential Revision: https://secure.phabricator.com/D8545
Summary: Fixes T912. This was very nearly working, it just needed a little tweaking on the last mile.
Test Plan:
Made updates with no effect, and updates with an effect. Made a no-effect update and posted just the comment part.
{F129037}
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T912
Differential Revision: https://secure.phabricator.com/D8543
Summary:
Fixes T3976. Long ago, some applications used "#comment-5" instead of "#5" for transaction/comment anchors. Now everything (I think?) uses "#5"; this is the style used by ApplicationTransactions.
This might break some very old, explcit `T123#comment-5` links, or off-site links to the `comment-N` anchors, but all that stuff generally got renumbered when we migrated anyway and getting you to the right object is like 95% of the job.
Test Plan: Verified that `T123#5` now links to `#5`.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T3976
Differential Revision: https://secure.phabricator.com/D8542
Summary: Fixes T4102. Document all the new stuff that CustomField supports now, and all the applications you can use it with.
Test Plan: Generated and read documentation.
Reviewers: chad, btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T4102
Differential Revision: https://secure.phabricator.com/D8541
Summary: This needs more updates after .arclint is ready to use, but should fix most of the biggest issues.
Test Plan: Generated and read it.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Differential Revision: https://secure.phabricator.com/D8539
Summary: Fixes T2791. I'm happy with HackerOne, so this pretty much just says "use HackerOne".
Test Plan:
{F128995}
- Clicked all the links.
Reviewers: btrahan, chad
Reviewed By: chad
Subscribers: epriestley
Maniphest Tasks: T2791
Differential Revision: https://secure.phabricator.com/D8538
Summary:
Fixes T3471. Specific issues:
- Add the ability to set a temporary cookie (expires when the browser closes).
- We overwrote 'phcid' on every page load. This creates some issues with browser extensions. Instead, only write it if isn't set. To counterbalance this, make it temporary.
- Make the 'next_uri' cookie temporary.
- Make the 'phreg' cookie temporary.
- Fix an issue where deleted cookies would persist after 302 (?) in some cases (this is/was 100% for me locally).
Test Plan:
- Closed my browser, reopned it, verified temporary cookies were gone.
- Logged in, authed, linked, logged out.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T3471
Differential Revision: https://secure.phabricator.com/D8537
Summary: Fixes T4430. Basically does a little code massage from the new stuff in D8525 and application transactions to get this working. Adds a new controller to the subscriptions app to make rendering these pretty easy peasy.
Test Plan: Used my test task in D8525 to verify both add and rem versions of these dialogs worked correctly.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: epriestley, chad, Korvin
Maniphest Tasks: T4430
Differential Revision: https://secure.phabricator.com/D8540
Summary:
Ref T2479, T4406. We should do a better job of (a) handling image processing errors and (b) declining to process large image files.
This fixes the worst of it, which is that users can upload huge GIFs with a large number of frames and hang a `convert` process for a long time, eating a CPU and a pile of memory.
This code is still pretty iffy and needs some more work. A near-term product goal for it is supporting 100x100 profile images.
Test Plan: Uploaded large and small GIFs, after setting the definition of "enormous" to be pretty small. Saw the small GIFs thumbnail into animated GIFs, and the large ones thumbnail into static images.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T2479, T4406
Differential Revision: https://secure.phabricator.com/D8536
Summary:
Fixes T4609. Steps are:
- Make a comment.
- Edit it.
- Delete all the text.
We expect to see "This comment has been deleted." -- instead, things currently render goofy.
Root cause is that `hasComment()` means both "comment object exists" //and// "comment object is nonempty".
Test Plan: {F128862}
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4609
Differential Revision: https://secure.phabricator.com/D8533
Summary:
Fixes T4610. Open to suggestions, etc., if there's anything I'm missing.
Also:
- Moves these "system" endpoints into a real application.
- Makes `isUnlisted()` work a little more consistently.
Test Plan: Accessed `/robots.txt`, `/status/` and `/debug/`.
Reviewers: chad, btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4610
Differential Revision: https://secure.phabricator.com/D8532
Summary:
Ref T4481. Summary is optional, but we currently always render it.
We previously rendered TEST PLAN. I wanted to see if anyone missed it. I miss it a little bit, and it sounds like @spicyj misses it. Restore it.
Test Plan:
$ ./bin/mail show-outbound --id 15232
...
BODY
epriestley created this revision.
epriestley added reviewers: The Bureaucracy, duck.
epriestley added a subscriber: duck.
TEST PLAN
more j
REVISION DETAIL
http://local.aphront.com:8080/D1042
AFFECTED FILES
number_j.txt
CHANGE DETAILS
Index: number_j.txt
===================================================================
--- number_j.txt
+++ number_j.txt
@@ -137,3 +137,4 @@
j
j
j
+j
To: epriestley, duck, Sebastiangarcia, Ahmedsmoore, nathanhthomas, chewnicorn
Cc: duck
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley, spicyj
Maniphest Tasks: T4481
Differential Revision: https://secure.phabricator.com/D8531
Summary: Fixes T4403. Supports the "send an email" action in Maniphest.
Test Plan: Wrote a "email duck" rule, then commented on a task and saw "duck" get an email.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4403
Differential Revision: https://secure.phabricator.com/D8529
Summary: Herald returns a map of `phid => true`. This is unintuitive and should probably be cleaned up eventually.
Test Plan: With a "Send an email to" rule, updated a revision and saw no error in error log.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8527
Summary: Fixes T4601. The "Differential Revision" field needs to be present in the "editable" version of the message so that `--verbatim` works correctly. Some day all of this might get rewritten to be a little easier to follow, maybe, but keep things working properly for now.
Test Plan: Used `arc diff`, `arc diff --edit`, `arc diff --verbatim`
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: aran, epriestley
Maniphest Tasks: T4601
Differential Revision: https://secure.phabricator.com/D8526
Summary: Ref T4430. This just deploys it on the property lists. (Help on how to do translations better? I tried a more traditional pht('%s, %s, %s, and %d other(s)') but I think the string lookup assumes the %d comes as the second param or something?)
Test Plan: Made a Maniphest Task with a hojillion subscribers and noted the working dialogue. Also made a Pholio Mock with lots of subscribers and it worked.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: aran, epriestley, Korvin, chad
Maniphest Tasks: T4430
Differential Revision: https://secure.phabricator.com/D8525
Summary: Fixes T4614. These don't do anything bad or dangerous, but generate unusable pages.
Test Plan:
- Added and executed unit tests.
- Tried to create pages like `/../`, `/begin/../end/`, etc.
Reviewers: btrahan, chad
Reviewed By: chad
Subscribers: aran, epriestley
Maniphest Tasks: T4614
Differential Revision: https://secure.phabricator.com/D8535
Summary: End-cap for timeline. Fixes T4438
Test Plan: Tested on a timeline with and without endcap.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: aran, epriestley, Korvin, chad, btrahan
Maniphest Tasks: T4438
Differential Revision: https://secure.phabricator.com/D8530
Summary:
I was a bit hasty with this.
- This should be uninstallable.
- Provide a real description.
- Choose a better title glyph (trident of neptune).
Test Plan: Poked around.
Reviewers: btrahan, chad
Reviewed By: chad
Subscribers: aran, epriestley
Differential Revision: https://secure.phabricator.com/D8534
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
