Summary: This allows the SavedQuery to modify what the result list looks like (e.g., include display flags and similar).
Test Plan: Looked at some ApplicationSearch apps.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2625
Differential Revision: https://secure.phabricator.com/D6346
Summary:
Ref T603. Ref T2625. Use cursors to page Differential queries, not offsets.
The trick here is that some queries are ordered. In these cases, we either need to pass some kind of tuple or do a cursor lookup. For example, if you are viewing revisions ordered by `dateModified`, we can either have the next page be something like:
?afterDateModified=2398329373&afterID=292&order=modified
...or some magical token:
?afterToken=2398329373:292&order=modified
I think we did this in Conpherence, but one factor there was that paging orders update with some frequency. In most cases, I think it's reasonable to pass just the ID and do a lookup to get the actual clause value (e.g., go look up object ID 292 and see what its dateModified is) and I think this is much simpler in general.
Test Plan: Set page size in Differential to 3, and paged through result lists ordered by date created and date modified.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603, T2625
Differential Revision: https://secure.phabricator.com/D6345
Summary:
Ref T603. Ref T2625. Makes `DifferentialRevisionQuery` do policy checks.
Note that it still uses inefficient offset-based paging, but it's rare to page through revisions. I'll switch to cursor paging in a future diff.
Test Plan: Viewed a bunch of Differential interfaces, home page, etc. This shouldn't actually materially impact anything.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603, T2625
Differential Revision: https://secure.phabricator.com/D6344
Summary:
Currently, when querying for responsible users (revisions where a given user is either the author or a reviewer), we do this:
- If the query passes a bunch of hard-coded special cases, use a special hard-coded UNION.
- Otherwise, use a very complicated JOIN/WHERE clause.
This is bad for several reasons:
- Tons and tons of hard-coding and special casing.
- The JOIN/WHERE clause performs very poorly for large datasets.
- (As a material consequence, the homepage issues a responsible query which barely misses the hard-coded special cases and goes down the slow path.)
Instead, //always// use the UNION strategy to execute a "responsible" query. Specifically, if we have responsible PHIDs, temporarily add them to the author list and build a normal query, then repeat for reviewers, then UNION any clauses we built.
Fixes T3377. Ref T603. Ref T2625. Depends on D6342.
There's various folklore about UNION ALL / UNION DISTINCT performance. UNION DISTINCT is simpler here and the number of rows is small, although we could use UNION ALL in the form:
SELECT * FROM ((SELECT ...) UNION ALL (SELECT ...) ORDER) GROUP LIMIT
...if we find that there's some performance benefit at some point.
Test Plan: Used DarkConsole to examine queries. Viewed home page and Differential dashboard/authors/subscribers.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603, T2625, T3377
Differential Revision: https://secure.phabricator.com/D6343
I don't actually have Phabricator installed locally so I have no idea if this works, but in theory this should add one of the + buttons in the left sidebar to quickly add a new macro.
Summary:
This fixes at least two issues with the S3 library on newer versions of cURL/PHP:
- NOTICE: PHP message: [2013-07-02 22:15:54] ERROR 8: curl_setopt(): CURLOPT_SSL_VERIFYHOST with value 1 is deprecated and will be removed as of libcurl 7.28.1. It is recommended to use value 2 instead at [/core/lib/phabricator/externals/s3/S3.php:1744]
- `$this->request->body` was appended to without initializing it, which rasies an error on PHP 5.5.0.
I looked over the rest of the changes briefly and they all seem reasonable-ish.
Test Plan:
- Uploaded a file to S3.
- Downloaded a file from S3.
- Deleted a file from S3.
- Checked error logs for anything suspicious.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D6349
Summary: This makes it namespace/database/connection aware and a little easier to find. Also use pht() / PhutilConsole.
Test Plan: Ran `bin/storage probe`.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D6341
Summary: They seem to look OK?
Test Plan: {F48529} {F48530}
Reviewers: chad
Reviewed By: chad
CC: aran
Differential Revision: https://secure.phabricator.com/D6350
Summary:
Ref T603. This introduces a policy-aware DifferentialDiffQuery and converts most callsites.
I've left unusual callsites (mostly: hard to get the viewer, unusual query, queries related to active diffs) alone for now, so this isn't exhaustive but hits 60-80% of sites.
Test Plan: Created diff; created revision; viewed diffs and revisions; made additional conduit calls.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D6338
Summary: Ref T603. This query isn't policy-aware yet, but prepare for it to be one day.
Test Plan: Looked at: home page; differential home; differential detail; diffusion browse. Made differential.query conduit call.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D6337
Summary: Ref T603. This is a very old, very bad version of `DifferentialRevisionQuery`. I want to modernize only the latter. Express the remaining callsite of the former in terms of `DifferentialRevisionQuery`.
Test Plan: Executed all four modes of `differential.find`.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D6335
Summary:
Ref T603.
- Primarily, this gets rid of a `DifferentialRevisionListData` callsite.
- Also modernize and clean up some UI stuff.
Test Plan:
{F48260}
{F48261}
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D6334
Summary:
- Add GC support to conduit logs.
- Add Query support to conduit logs.
- Record the actual user PHID.
- Show client name.
- Support querying by specific method, so I can link to this from a setup issue.
@wez, this migration may not be fast. It took about 8 seconds for me to migrate 800,000 rows in the `conduit_methodcalllog` table. This adds a GC which should keep the table at a more manageable size in the future.
You can safely delete all data older than 30 days from this table, although you should do it by `id` instead of `dateCreated` since there's no key on `dateCreated` until this patch.
Test Plan:
- Ran GC.
- Looked at log UI.
- Ran Conduit methods.
Reviewers: btrahan
Reviewed By: btrahan
CC: wez, aran
Differential Revision: https://secure.phabricator.com/D6332
Summary:
Ref T603. Ref T2625.
Long chain of "doing the right thing" here: I want to clean this up, so I can clean up the Conduit logs, so I can add a setup issue for deprecated method calls, so I can remove deprecated methods, so I can get rid of `DifferentialRevisionListData`, so I can make Differntial policy-aware.
Adds modern infrastructure and UI to all of the Conduit interfaces (except only partially for the logs, that will be the next diff).
Test Plan:
{F48201}
{F48202}
{F48203}
{F48204}
{F48206}
This will get further updates in the next diff:
{F48205}
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T603, T2625
Differential Revision: https://secure.phabricator.com/D6331
Summary:
Depends on D6329. This fixes `http://www.example.com/D123`, which currently gets the "D123" rendered, after addition of the Asana rule. It also removes a hack for object refernces.
Basically, the "hyperlink" rule needs to happen after rules which specialize hyperlinks (Youtube, Asana) but before rules which apply to general text (like the Differential and Maniphest rules). Allow these rules to specify that they have higher or lower priority.
Test Plan: Asana rules, Differential rules and Diffusion rules now all markup correctly.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D6330
Summary: See discussion in IRC. Not 100% sure what's going on here because of email ghost theives, but conceivably a commit with no changes will end up with `null` changesets instead of `array()` changesets, which throws. Such diffs are certianly possible (`git commit --allow-empty`) even if they aren't the issue in this specific case. See T3416. Initialize changesets to `array()` to avoid throwing.
Test Plan:
Viewed some commits?
iiam
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Differential Revision: https://secure.phabricator.com/D6339
Summary:
Addes a button group to filter tasks by agents, non-agents or all.
Fixes T3394
Test Plan: View task list, filter by agents, filter by non agents. Make sure the correct tasks display.
Reviewers: epriestley, dctrwatson
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T3394
Differential Revision: https://secure.phabricator.com/D6328
Summary: Fixes T3400. Users are crafty. Attempt to outwit them.
Test Plan: Added all kinds of nonsense to my PATH to hit all the errors. Verified sensible-looking error messages which I couldn't figure out any way to misread or outwit.
Reviewers: chad, btrahan
Reviewed By: chad
CC: aran
Maniphest Tasks: T3400
Differential Revision: https://secure.phabricator.com/D6318
Summary:
Ref T3116. This is a large amount of schema for V0 but it seems relatively complete to the desired features in T3116.
The only thing of note that is missing is documentSignatures should have some sort of "signedStatus". "Un-signing" seemed weird to me, though I could imagination "pending signature". "Pending signature" could be done via edges pretty easily.
Plan is to have "Document" be at the top level and own policy. "DocumentBody" will store a version of title and text for each and every "edit" on a larger Document. "Edges" are to be used to tie Authors => Document for V0ish. Transactions are going to be used to store all the various edits possible here. Oh and DocumentSignatures will do what you expect, but include documentVersion as part of the key.
Test Plan: just some schema. `storage update` worked though!
Reviewers: epriestley
Reviewed By: epriestley
CC: chad, aran, Korvin
Maniphest Tasks: T3116
Differential Revision: https://secure.phabricator.com/D6323
Summary: I have simply copied the existing css into the oblivious skin. I don't know if this is the right approach (code duplication), but considering this skin should be isolated (and will potentially differ) I think this makes sense.
Test Plan: Use a code block on a phame post.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin, chad
Differential Revision: https://secure.phabricator.com/D6313
Summary:
Fixes T3436. Currently, when installs have configuration options which we don't know about, we raise a fairly confusing/ambiguous message about the options being unknown. Instead:
- Keep a list of previously valid (but now deleted) config, with explanatory reasons for what happened to it. Present this information, along with altenate wording ("Obsolete Config" instead of "Unknown Config") where applicable.
- Show a list of all the places the config is defined.
- Provide an active link to delete it from the web UI.
- Provide a command to delete it from the CLI.
- Allow `bin/config delete` to delete configuration options which no longer have a definition.
Test Plan:
- Set an auth key in database, local and file config.
- Walked through the setup issue, cleaning it up.
- Set an invalid key and made sure I still got a reasonable error (this now has better cleanup instructions).
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T3436
Differential Revision: https://secure.phabricator.com/D6317
Summary: Currently you can't refresh the live blog or a blog post after clicking "View Live" due to POST action. I have removed the setRenderAsForm call on the "View Live" actions. I am unsure if this has any unintended consequences but I have tested and not found any.
Test Plan: Click the "View Live" action within a blog post or blog, and observe that the request occurs via GET not POST
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D6314
Summary: Ref T2852. Setting followers (like CCs) is a separate API call, but we don't need to do anything complicated.
Test Plan: Synchronized revisions and verified the parent task got followers.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6308
Summary:
Ref T2852.
Before trying related users, try using the feed story's actor. This is the most correct voice to act in.
Test Plan: Ran `feed/republish`.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6305
Summary:
Ref T2852.
The parent task is open unless the revision is in the states "closed" or "abandoned". If it's in "needs review", it remains open. This last bit is slightly unlike Differential, but consistent with the Google Doc and generally seems like a better fit. There's no way to put the task in a "Waiting on Others" state in Asana like we can in Differential.
The subtasks are closed unless the revision is in the state "needs review". This is generally consistent with Differential.
Test Plan:
Made a series of changes to a revision and synchronized it repeatedly:
- requested changes
- commandeered
- requested review
- abandoned
Verified task and subtasks synchronized states correctly in Asana.
{F47554}
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6304
Summary:
Ref T2852. Depends on D6302. This now creates, destroys, and synchronizes subtasks.
- After finishing the parent task stuff, we pull a list of all known subtasks.
- We load all those subtasks.
- If we fail to load any, we delete their objects and edges on the Phabricator side.
- Of the remaining subtasks, we find subtasks for users who aren't related to the object any more and delete them in Asana and locally (for example, if alincoln is removed as a reviewer, we delete his subtask).
- For all the related users, we either synchronize their existing task or create a new one for them.
- Then we write edges for any new tasks we added.
This doesn't handle a few weird edge cases in any specific way:
- If a subtask is moved under a different parent, we ignore it.
- If a new subtask is created that we don't know about, we ignore it.
- If a subtask we know about is deleted, we just respawn it. This is consistent with "DON'T EDIT THESE". You can force sync to stop by deleting the parent.
Addititionally:
- Make the "don't edit" warning more compelling and visceral.
Test Plan:
- Kind of ran it a bit.
- There are like 3,000 edge cases here so this is hard to test exhaustively.
- Forced a few of the edge cases to happen.
- Nothing seems immediately broken in an obvious way?
{F47551}
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6303
Summary:
Ref T2852. This is highly incomplete but seems structurally sound. Some additional context is available in the Google doc.
- Add a workspace ID configuration. Without it, nothing else activates.
- Add a worker which reacts to feed stories.
- Feed stories about things which aren't Differential objects are ignored.
- We load the revision, or fail permanently if we can't.
- We get all the related user PHIDs (author, reviewers, CCs).
- We check if any of them have linked Asana accounts, or fail permanently if they don't.
- We check for an "ASANATASK" edge from the revision.
- If we do not find one, we create a new task.
- If we do find one, we load the task.
- If we succeed, we check the chronological key of the most recent synchronized feed story ("cursor").
- If this story is the same or newer, we update the task to synchronize it to the current state of the revision.
- If we fail to load the task, we fail permanently ("asana task has been deleted").
- We then publish the actual story text to the task.
Not in yet:
- Updating followers requires separate API calls which we don't do yet.
- No subtasks yet.
- No sync of open/closed state.
Test Plan: {F47546}
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6302
Summary: Currently if disqus is used and a phame post is loaded over ssl, the disqus api is not loaded over https. This fixes that by honouring the protocol being used by the html document.
Test Plan: Open a phame post over https
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D6311
Summary: Ref T2852. Reduce the number of magical strings in use, and prepare the Asana bridge for eventual workspace/project support (a little bit).
Test Plan: Verified enriched links still work properly.
Reviewers: btrahan, chad
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6301
Summary:
Ref T2852. Add a `log()` method to `PhabricatorWorker` to make debugging easier.
I renamed the similar Drydock-specific method.
Test Plan:
Used logging in a future revision:
...
<<< [36] <http> 211,704 us
Updating main task.
>>> [37] <http> https://app.asana.com/api/1.0/tasks/6153776820388
...
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6296
Summary:
Ref T1536. Ref T2852. Currently, after refreshing the token we don't actually return it. This means that code relying on token refresh fails once per hour (for Asana) in a sort of subtle way. Derp.
Update `bin/auth refresh` to make this failure more clear.
Test Plan: Set `force refresh` flag and verified a return value.
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T1536, T2852
Differential Revision: https://secure.phabricator.com/D6295
Summary:
Ref T2852.
- Broadly, we support "I have a Ref, I need a PHID" well but not "I have a PHID, I need a Ref".
- Add DoorkeeperExternalObjectQuery, and use it to query ExternalObjects.
- Allow external objects to be imported by their internal PHIDs. Basically, if we have an edge pointing at an ExternalObject, we can say "load all the data about this" from just the PHID and have it hit all the same code.
- Allow construction of Refs from ExternalObjects. This makes the "I have a PHID, I need a Ref" easier.
Test Plan:
- Verified Asana links still enrich properly at display time.
- Used in future revision.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6294
Summary:
Ref T2852. When we pull edge data and an edge has none, we currently populate `null` in the results. This is inconvenient and makes `idx()`'ing it clumsy. Instead, populate `array()` for empty.
(We've barely used edge data anywhere so far, which is why this hasn't come up before, but I have some use cases for it now.)
Test Plan:
- Trivial / used in future diff.
- Verified existing edge data callsites don't care about this API change (there are only 3).
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6293
Summary:
Ref T2852. I want to model Asana integration as a response to feed events. Currently, we queue one feed event for each HTTP hook.
Instead, always queue one feed event and then have it queue any necessary followup events (now, http hooks; soon, asana).
Add a script to make it easy to reproducibly fire feed event publishing.
Test Plan:
Republished a feed event and verified it hit configured HTTP hooks correctly.
$ ./bin/feed republish 5765774156541908292 --trace
>>> [2] <connect> phabricator2_feed
<<< [2] <connect> 1,660 us
>>> [3] <query> SELECT story.* FROM `feed_storydata` story JOIN `feed_storyreference` ref ON ref.chronologicalKey = story.chronologicalKey WHERE (ref.chronologicalKey IN (5765774156541908292)) GROUP BY story.chronologicalKey ORDER BY story.chronologicalKey DESC
<<< [3] <query> 595 us
>>> [4] <connect> phabricator2_differential
<<< [4] <connect> 760 us
>>> [5] <query> SELECT * FROM `differential_revision` WHERE phid IN ('PHID-DREV-ywqmrj5zgkdloqh5p3c5')
<<< [5] <query> 478 us
>>> [6] <query> SELECT * FROM `differential_revision` WHERE phid IN ('PHID-DREV-ywqmrj5zgkdloqh5p3c5')
<<< [6] <query> 449 us
>>> [7] <connect> phabricator2_user
<<< [7] <connect> 1,062 us
>>> [8] <query> SELECT * FROM `user` WHERE phid in ('PHID-USER-lqiz3yd7wmk64ejugvov')
<<< [8] <query> 540 us
>>> [9] <connect> phabricator2_file
<<< [9] <connect> 951 us
>>> [10] <query> SELECT * FROM `file` WHERE phid IN ('PHID-FILE-gq6dlsysvxbn3dgwvky7')
<<< [10] <query> 498 us
>>> [11] <query> SELECT * FROM `user_status` WHERE userPHID IN ('PHID-USER-lqiz3yd7wmk64ejugvov') AND UNIX_TIMESTAMP() BETWEEN dateFrom AND dateTo
<<< [11] <query> 507 us
Republishing story...
>>> [12] <query> SELECT story.* FROM `feed_storydata` story JOIN `feed_storyreference` ref ON ref.chronologicalKey = story.chronologicalKey WHERE (ref.chronologicalKey IN (5765774156541908292)) GROUP BY story.chronologicalKey ORDER BY story.chronologicalKey DESC
<<< [12] <query> 685 us
>>> [13] <query> SELECT * FROM `differential_revision` WHERE phid IN ('PHID-DREV-ywqmrj5zgkdloqh5p3c5')
<<< [13] <query> 489 us
>>> [14] <query> SELECT * FROM `differential_revision` WHERE phid IN ('PHID-DREV-ywqmrj5zgkdloqh5p3c5')
<<< [14] <query> 512 us
>>> [15] <query> SELECT * FROM `user` WHERE phid in ('PHID-USER-lqiz3yd7wmk64ejugvov')
<<< [15] <query> 601 us
>>> [16] <query> SELECT * FROM `file` WHERE phid IN ('PHID-FILE-gq6dlsysvxbn3dgwvky7')
<<< [16] <query> 405 us
>>> [17] <query> SELECT * FROM `user_status` WHERE userPHID IN ('PHID-USER-lqiz3yd7wmk64ejugvov') AND UNIX_TIMESTAMP() BETWEEN dateFrom AND dateTo
<<< [17] <query> 551 us
>>> [18] <query> SELECT story.* FROM `feed_storydata` story JOIN `feed_storyreference` ref ON ref.chronologicalKey = story.chronologicalKey WHERE (ref.chronologicalKey IN (5765774156541908292)) GROUP BY story.chronologicalKey ORDER BY story.chronologicalKey DESC
<<< [18] <query> 507 us
>>> [19] <query> SELECT * FROM `differential_revision` WHERE phid IN ('PHID-DREV-ywqmrj5zgkdloqh5p3c5')
<<< [19] <query> 428 us
>>> [20] <query> SELECT * FROM `differential_revision` WHERE phid IN ('PHID-DREV-ywqmrj5zgkdloqh5p3c5')
<<< [20] <query> 419 us
>>> [21] <query> SELECT * FROM `user` WHERE phid in ('PHID-USER-lqiz3yd7wmk64ejugvov')
<<< [21] <query> 591 us
>>> [22] <query> SELECT * FROM `file` WHERE phid IN ('PHID-FILE-gq6dlsysvxbn3dgwvky7')
<<< [22] <query> 406 us
>>> [23] <query> SELECT * FROM `user_status` WHERE userPHID IN ('PHID-USER-lqiz3yd7wmk64ejugvov') AND UNIX_TIMESTAMP() BETWEEN dateFrom AND dateTo
<<< [23] <query> 593 us
>>> [24] <http> http://127.0.0.1/derp/
<<< [24] <http> 746,157 us
[2013-06-24 20:23:26] EXCEPTION: (HTTPFutureResponseStatusHTTP) [HTTP/500] Internal Server Error
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6291
Summary: Currently setting "disqus.shortname" via config isn't actually used in the instantiation of disqus. This fix uses the shortname configured.
Test Plan: Open a phame post with disqus enabled and a shortname other than "phabricator" specified.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D6310
Summary: When I ignore setup issues, I want them to look dealt with, and keep yellow for new ones. Also updated callout colors.
Test Plan: Ignored a number of issues.
Reviewers: epriestley, btrahan
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D6300
Summary: Ref T3322. Depends on D6297. Here are some Phabricator tweaks to complment D6297.
Test Plan: {F47522}
Reviewers: garoevans
Reviewed By: garoevans
CC: aran, chad
Maniphest Tasks: T3322
Differential Revision: https://secure.phabricator.com/D6298
Summary:
Ref T1536. This is equivalent to logging out and logging back in again, but a bit less disruptive for users. For some providers (like Google), this may eventually do something different (Google has a "force" parameter which forces re-auth and is ostensibly required to refresh long-lived tokens).
Broadly, this process fixes OAuth accounts with busted access tokens so we can do API stuff. For other accounts, it mostly just syncs profile pictures.
Test Plan:
Refreshed LDAP and Oauth accounts, linked OAuth accounts, hit error conditions.
{F47390}
{F47391}
{F47392}
{F47393}
{F47394}
{F47395}
Reviewers: btrahan, chad
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1536
Differential Revision: https://secure.phabricator.com/D6290
Summary:
Ref T1536.
- Allow providers to customize the look of external accounts.
- For username/password auth, don't show the account view (it's confusing and not useful).
- For OAuth accounts, show token status.
Test Plan:
{F47374}
{F47375}
Reviewers: btrahan, chad
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1536
Differential Revision: https://secure.phabricator.com/D6289
Summary:
Ref T2852. Give OAuth providers a formal method so you can ask them for tokens; they issue a refresh request if necessary.
We could automatically refresh these tokens in daemons as they near expiry to improve performance; refreshes are blocking in-process round trip requests. If we do this for all tokens, it's a lot of requests (say, 20k users * 2 auth mechanisms * 1-hour tokens ~= a million requests a day). We could do it selectively for tokens that are actually in use (i.e., if we refresh a token in response to a user request, we keep refreshing it for 24 hours automatically). For now, I'm not pursuing any of this.
If we fail to refresh a token, we don't have a great way to communicate it to the user right now. The remedy is "log out and log in again", but there's no way for them to figure this out. The major issue is that a lot of OAuth integrations should not throw if they fail, or can't reasonably be rasied to the user (e.g., activity in daemons, loading profile pictures, enriching links, etc). For now, this shouldn't really happen. In future diffs, I plan to make the "External Accounts" settings page provide some information about tokens again, and possibly push some flag to accounts like "you should refresh your X link", but we'll see if issues crop up.
Test Plan: Used `bin/auth refresh` to verify refreshes. I'll wait an hour and reload a page with an Asana link to verify the auto-refresh part.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6280
Summary: Ref T2852. Provide a script for inspecting/debugging OAuth token refresh.
Test Plan: Ran `bin/auth refresh` with various arguments, saw token refreshes.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6276
Summary:
Ref T2852. Primarily, this expands API access to Asana. As a user-visible effect, it links Asana tasks in Remarkup.
When a user enters an Asana URI, we register an onload behavior to make an Ajax call for the lookup. This respects privacy imposed by the API without creating a significant performance impact.
Test Plan: {F47183}
Reviewers: btrahan
Reviewed By: btrahan
CC: chad, aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6274
