Summary: Fixes T2328. Note the audit part is fixed now.
Test Plan: Tried to reproduce the audit issue by raising my own commit as a problem; it showed up before code changes! Made a diff with my self as author and reviewer; it showed up as expected
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: epriestley, Korvin
Maniphest Tasks: T2328
Differential Revision: https://secure.phabricator.com/D8755
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. Update this callsite; pretty straightforward.
Test Plan: Used Conduit to take actions and saw their effects in Differential.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8442
Summary:
Ref T3496. Currently, we call loadAssets() on each revision table, which invokes a new revision query and a pile of subqueries.
Instead, add `needFlags()` and `needDrafts()` to `RevisionQuery`. Some day these could perhaps be more generic.
Test Plan:
- Viewed home, differential, etc., no longer saw 9203809238 queries being run for no reason.
- Drafts and flags still appear properly.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T3496
Differential Revision: https://secure.phabricator.com/D8277
Summary: ...do it somewhat generically, so we could fairly easily add this to other applications. Fixes T3496. I got a wee bit lazy and decided not to migrate existing drafts. My excuses aside from laziness are doing it this way will let us see if anyone complains, we can always do a migration later if people do complain, and there's likely to be a lot of garbage data for older / bigger installs, and the migration didn't seem worth itgiven it would also likely be expensive in these cases.
Test Plan: made a draft inline comment on DX and observed DX had a note icon on Differential home page. made a draft comment on DX and observed DX had a note icon on Differential home page. deleted a draft inline comment and noted icon disappeared from Differential homepage. Submitted a draft comment + inline comment and noted icon disappeared.
Reviewers: epriestley
Reviewed By: epriestley
CC: Korvin, epriestley, aran
Maniphest Tasks: T3496
Differential Revision: https://secure.phabricator.com/D8275
Summary:
Ref T2222. I want to stage a "later" patch to drop this column, but get rid of the last few references to it.
One of these methods has no callers, and the other stuff I've updated to use the modern fields.
Test Plan: Created some inlines, hit "edit", submitted them, `grep.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8240
Summary:
Ref T2222. This is the big one.
This migrates each `DifferentialComment` to one or more ApplicationTransactions (action, cc, reviewers, update, comment, inlines), and makes `DifferentialComment` a double-reader for ApplicationTransactions.
The migration is pretty straightforward:
- If a comment took an action not otherwise covered, it gets an "action" transaction. This is something like "epriestley abandoned this revision.".
- If a comment updated the diff, it gets an "updated diff" transaction. Very old transactions of this type may not have a diff ID (probably only at Facebook).
- If a comment added or removed reviewers, it gets a "changed reviewers" transaction.
- If a comment added CCs, it gets a "subscribers" transaction.
- If a comment added comment text, it gets a "comment" transaction.
- For each inline attached to a comment, we generate an "inline" transaction.
Most comments generate a small number of transactions, but a few generate a significant number.
At HEAD, the code is basically already doing this, so comments in the last day or two already obey these rules, roughly, and will all generate only one transaction (except inlines).
Because we've already preallocated PHIDs in the comment text table, we only need to write to the transaction table.
NOTE: This significantly degrades Differential, making inline comments pretty much useless (they each get their own transaction, and don't show line numbers or files). The data is all fine, but the UI is garbage now. This needs to be fixed before we can deploy this to users, but it's easily separable since it's all just display code.
Specifically, they look like this:
{F112270}
Test Plan:
I've migrated locally and put things through their paces, but it's hard to catch sketchy stuff locally because most of my test data is nonsense and bad migrations wouldn't necessarily look out of place.
IMPORTANT: I'm planning to push this to a branch and then shift production over to the branch, and run it for a day or two before bringing it to master.
I generally feel good about this change: it's not that big since we were able to separate a lot of pieces out of it, and it's pretty straightforward. That said, it's still one of the most scary/dangerous changes we've ever made.
Reviewers: btrahan
CC: chad, aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8210
Summary:
Ref T2222. Ref T4415. We're still writing Differential subscription stuff into this weird legacy `differential_relationship` table, which is like an edge table but extremely ancient.
Move it into a proper table.
I've removed `withSubscriptions()` from `DifferentialRevisionQuery`. It was weird, doesn't work consistently with other similar filters, and was only used by the API. Now it means "ccs", which is consistent with the ApplicationSearch UI and with Maniphest.
Test Plan:
Without migrating, added and removed subscribers via various workflows. Queried for subscribers. Everything worked as expected.
Ran the migration, verified data survived.
Reviewers: btrahan
Reviewed By: btrahan
CC: FacebookPOC, aran
Maniphest Tasks: T2222, T4415
Differential Revision: https://secure.phabricator.com/D8202
Summary:
Ref T2222. Currently, `DifferentialComment` stores both (a) the text of comments and (b) various other transaction details. This data needs to map to both Transactions and TransactionComments in the long run. This diff separates out all the data which is bound for the TransactionComment table, so that when we migrate `DifferentialComment` itself it will //only// need to migrate into the Transactions table. This is a much simpler migration than the inline comment one was, partly because it set up infrastructure and partly because the data is less complex.
Basically, I'm just proxying the read/write for the comment text into the other table. All readers already go through the Query class, and there are only three writers (preview, comment, implicit comment on diff update) which are all highly regular and straightforward to test.
We can also back out of this diff very easily: doing double writes cost only one line of code (`$this->content = $content;`) so we have proper double writes and a trivial revert path.
Test Plan:
- Without migrating, added comments and saw them show up.
- Migrated.
- Saw all the old comments, and no damage to the new ones.
- Added new comments.
- Used comment preview.
- Updated a revision to implicitly create an update comment and verified it looked OK.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D8196
Summary:
Currently, "Closed" and "Abandoned" are treated as "closed". I want to add a flag which treats "Accepted" as "Closed", too, for Asana and other companies who use an Asana-like workflow.
The background here is that their workflow is a bit weird. They basically do audits, but have a lot of things which Diffusion doesn't do well right now. This one change makes Differential fit their workflow fairly well, even though it's an audit workflow.
To prepare for this, normalize the definition of "closed" better. We have a few callsites which explicitly check for "ABANDONED || CLOSED", and normalizing this is cleaner anyway.
Also delete the very old COMMITTED status, which has been obsolete for over a year.
Test Plan: Browsed around most/all of the affected interfaces.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D7653
Summary:
Fixes T4109. If a revision has a bad `repositoryPHID` (for example, because the repository was deleted), `DifferentialRevisionQuery` calls `didRejectResult()` on it, which raises a policy exception, even if the viewer is omnipotent. This aborts the `MessageParser`, because it does not expect policy exceptions to be raised for an omnipotent viewer.
Fix this in two ways:
# Never raise a policy exception for an omnipotent viewer. I think this is the expected behavior and a reasonable rule.
# In this case, load the revision for an omnipotent viewer.
This feels a little gross, but it's the only place where we do this in the codebase right now. We can clean this up later on once it's more clear what the circumstances of checks like these are.
Test Plan: Set a revision to have an invalid `repositoryPHID`, ran message parser on it, got a clean parse.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4109
Differential Revision: https://secure.phabricator.com/D7603
Summary:
Ref T1049. Ref T2222. `DifferentialDiff` does not currently have a PHID, but we need it for Harbormaster and ApplicationTransactions. See some discussion in D7501.
(I split the SQL into two sections so we can't fail in the middle. At some point, I'd like to do a pass on the migration stuff and get this happening automatically, and also simplify the PatchList.)
Test Plan:
- Ran `bin/storage upgrade`.
- Checked for valid PHIDs in the database.
- Used `phid.query` to look up a diff by PHID.
- Created a new diff and verified it got a PHID.
Reviewers: btrahan, hach-que
Reviewed By: btrahan
CC: aran, vrana
Maniphest Tasks: T2222, T1049
Differential Revision: https://secure.phabricator.com/D7513
Summary:
While we mostly have reasonable effective object accessibility when you lock a user out of an application, it's primarily enforced at the controller level. Users can still, e.g., load the handles of objects they can't actually see. Instead, lock the queries to the applications so that you can, e.g., never load a revision if you don't have access to Differential.
This has several parts:
- For PolicyAware queries, provide an application class name method.
- If the query specifies a class name and the user doesn't have permission to use it, fail the entire query unconditionally.
- For handles, simplify query construction and count all the PHIDs as "restricted" so we get a UI full of "restricted" instead of "unknown" handles.
Test Plan:
- Added a unit test to verify I got all the class names right.
- Browsed around, logged in/out as a normal user with public policies on and off.
- Browsed around, logged in/out as a restricted user with public policies on and off. With restrictions, saw all traces of restricted apps removed or restricted.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D7367
Summary:
Ref T2222. This implements step (1) described there, which is moving over all the inline comments.
The old and new tables are simliar. The only real trick here is that `transactionPHID` and `legacyCommentID` mean roughly the same thing (`null` if the inline is a draft, non-null if it has been submitted) but we don't have real `transactionPHID`s yet. We just make some up -- we'll backfill them later.
Two risks here:
- I need to take a second look at the keys on this table. I think we need to tweak them a bit, and it will be less disruptive to do that before this migration than after.
- This will take a while for Facebook, and other large installs with tens of thousands of revisions. I'll communicate this.
I'm otherwise pretty satisfied with this, seems to work well and is pretty low risk / non-disruptive.
Test Plan:
- Before migrating, then after migrating:
- Made a bunch of inlines (drafts, submitted).
- Edited and deleted inlines.
- Verified inlines showed up in preview.
- Verified that inlines aren't indexed when they're drafts (`bin/search index D935`).
- Verified that inlines ARE indexed when they're not drafts.
- Verified that drafts inlines make revisions appear as "with draft" in the revision list.
- Made left, right, and draft inlines.
- Migrated (`bin/storage upgrade`).
- Verified that my inlines from before the migration still showed up.
- (Repeated all the stuff above.)
- Manually inspected the inline comment table.
Reviewers: btrahan
Reviewed By: btrahan
CC: FacebookPOC, aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D7139
Summary:
Ref T1279. We currently determine reviewers at display time, but this is bad for several reasons:
- It puts queries very close to the display layer.
- We have to query for each revision if we want to figure out authority for several.
- We need to figure it out in several places, so we'll end up with copies of this logic.
- The logic isn't trivial (exceptions for the viewer, exceptions to that rule for install configuration).
- We already do this "figure it out when we need it" stuff in Diffusion for audits and it's really bad: we have half-working copies of the logic spread all over the place.
Instead, put it in the Query. Callers query for it and get the data attached to the reviewer objects.
Test Plan:
- Looked at some revisions, verified the correct lines were highlighted.
- Looked at a revision I created and verified that projects I was a member of were not highlighted.
- With self-accept enabled, these //are// highlighted.
- Looked at a revision I did not create and verified that projects I was a member of were highlighted.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1279
Differential Revision: https://secure.phabricator.com/D7241
Summary:
Ref T1279. Two changes to the search/query for Differential:
- "Reviewers" now accepts users and projects.
- "Responsible Users" now includes revisions where a project you are a member of is a reviewer.
Test Plan:
- Searched for project reviewers.
- Verified that the dashboard now shows reviews which I'm only part of via project membership.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1279
Differential Revision: https://secure.phabricator.com/D7231
Summary:
Ref T1279. No actual logical changes, but:
- You can now add projects as reviewers from the revision view typeahead ("Add Reviewers" action).
- You can now add projects as reviewers from the revision detail typeahead.
- You can now add projects as reviewers from the CLI (`#yoloswag`).
- Generated commit messages now list project reviewers (`Reviewers: #yoloswag`).
I'll separate projects from users in the "Reviewers" tables in the next revision.
Test Plan:
- Added projects as reviewers using the web UI and CLI.
- Used `arc amend --show --revision Dnnn` to generate commit messages.
- Viewed revision with project reviewers in web UI.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1279
Differential Revision: https://secure.phabricator.com/D7230
Summary:
Ref T1279. @champo did a lot of this work already; we've been doing double writes for a long time.
Add "double reads" (reading the edge table as both the "relationship" table and as the "reviewer status" table), and migrate all the data.
I'm not bothering to try to recover old reviewer status (e.g., we could infer from transactions who accepted old revisions) because it wold be very complicated and doesn't seem too valuable.
Test Plan:
- Without doing the migration, used Differential. Verified that reads and writes worked. Most of the data was there anyway since we've been double-writing.
- Performed the migration. Verified that everything was still unchanged.
- Dropped the edge table, verified all reviweer data vanished.
- Migrated again, verified the reviewer stuff was restored.
- Did various cc/reviewer/subscriber queries, got consistent results.
Reviewers: btrahan
Reviewed By: btrahan
CC: champo, aran
Maniphest Tasks: T1279
Differential Revision: https://secure.phabricator.com/D7227
Summary:
Ref T603. This could probably use a little more polish, but improve the quality of policy error messages.
- Provide as much detail as possible.
- Fix all the strings for i18n.
- Explain special rules to the user.
- Allow indirect policy filters to raise policy exceptions instead of 404s.
Test Plan: See screenshots.
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D7151
Summary:
Ref T603. Make Differential behaviors for logged-out and underprivleged users more similar to other apps.
I'm going to drop this "anonymous access" thing at some point, but `reviews.fb.net` actually looks like it's running semi-modern code, so leave it alive until we have a more compelling replacement in the upstream.
Test Plan: As a logged out user, browsed Differential and clicked things and such.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D7148
Summary:
Ref T603. When a diff is attached to a revision, try to guess the repository if possible. In cases where we succeed, this automatically gives us intuitive policy behavior (i.e., you can see a revision if you can see the repository the change is against).
I pulled this into a funky little "Lookup" class for two reasons:
- It's used in two places;
- I anticipate that we might need to add some sort of `explainWhy()` method if users find the heuristics confusing.
Test Plan: Created and updated revisions, saw them pick up the correct repository association. Ran Herald dry run against associable and nonassociable revisions, saw correct values populate.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D7147
Summary: This isn't too useful most of the time since we don't automatically populate this data yet, but works fine.
Test Plan: See screenshot.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D7144
Summary:
Ref T603. Read policies out of policy columns.
When a revision is associated with a repository (which is currently never), require view access on the repository to see the revision (or, require the viewer to be the owner). This is a blanket "do the right thing" rule which should make Differential's default policies align with user expectations.
Future diffs will populate the `repositoryPHID` when a revision is created.
Test Plan: Tooled around Differential. None of this stuff does anything yet, so nothing very exciting happened.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D7134
Summary:
`getArcanistProjectName()` has some logic which gets messy with the `self::ATTACHABLE` mechanism. This makes `differential.getdiff` and similar Conduit methods throw an exception when querying a diff which doesn't have a project. See <http://pastebin.com/Czzrd0Jz>.
Instead, unconditionally attach a project (possibly `null`) when loading diffs if they need projects.
Test Plan: Ran `differential.getdiff` against a `arc diff --raw` diff with no project, got a result instead of an exception.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran, sttwister
Differential Revision: https://secure.phabricator.com/D7101
Summary:
Ref T603. See inlines for an explanation. The case where I hit this was loading the "Pending Differential Revisions" panel in Diffusion when logged out, after making a repository public.
What happens is that we load 10 revisions (say, D1 .. D10) but the user can't see any of them. We then try to load the next 10, but since the pagination is ordered by date modified, we need to base the next query on the modified date of the last thing we loaded (D10). However, since we use the viewer's policies to load that cursor object, it fails to load, and then we just issue the same query over and over again, loading D1 .. D10 until we run out of execution time.
Test Plan: Interface now loads correctly.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D7059
Summary: I kind of made a mess of the API doing T2784. I figure just adding this is fine but LMK if you'd prefer something like diffquery got cleaned up more to handle this. Also adds an idx() call as I was getting errors looking at old diffs. Fixes T3823.
Test Plan: used the new api via test console - great success.
Reviewers: epriestley
Reviewed By: epriestley
CC: Korvin, aran
Maniphest Tasks: T3823
Differential Revision: https://secure.phabricator.com/D6966
Summary:
We currently have two giant messes for paging across multiple columns (usually because one column is not unique), and I'm about to add a third for Maniphest.
Provide a more structured way to build these `A > a OR (A = a AND B > b)` clauses.
Test Plan: Set page size to `2` for Differential and Diffusion and paged forward and backward with a bunch of different orders set. Pages worked as expected.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2625
Differential Revision: https://secure.phabricator.com/D6971
Summary: Ref T603. Killing this class is cool because the classes that replace it are policy-aware. Tried to keep my wits about me as I did this and fixed a few random things along the way. (Ones I remember right now are pulling a query outside of a foreach loop in Releeph and fixing the text in UIExample to note that the ace of hearts if "a powerful" card and not the "most powerful" card (Q of spades gets that honor IMO))
Test Plan: tested the first few changes (execute, executeOne X handle, object) then got real mechanical / careful with the other changes.
Reviewers: epriestley
Reviewed By: epriestley
CC: Korvin, aran, FacebookPOC
Maniphest Tasks: T603
Differential Revision: https://secure.phabricator.com/D6941
Summary: Fixes T3786. Not 100% sold on this (I don't want to restore all of the original filters, since users can and should just build the weird ones if they use them), but this is almost certainly the most useful of the defaults which ApplicationSearch removed.
Test Plan: Viewed `/differential/`, executed the query.
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T3786
Differential Revision: https://secure.phabricator.com/D6860
Summary:
Fixes T3781. The UI defaults to "Created" but the query defaults to "Modified". Make the two consistent.
In particular, an issue this fixes is that previously a `/differential/?authors=duck` page would show "Order: Created" but actually order by "Modified".
Test Plan: Visited `/differential/?authors=duck` and verified the revisions were ordered by creation date.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T3781
Differential Revision: https://secure.phabricator.com/D6843
Summary:
Ref T3775 (discussion here). Ref T2625.
T3775 presents two problems:
# Existing tools which linked to `/differential/active/epriestley/` (that is, put a username in the URL) can't generate search links now.
# Humans can't edit the URL anymore, either.
I think (1) is an actual issue, and this fixes it. I think (2) is pretty fluff, and this doesn't really try to fix it, although it probably improves it.
The fix for (1) is:
- Provide a helper to read a parameter containing either a list of user PHIDs or a list of usernames, so `/?users[]=PHID-USER-xyz` (from a tokenizer) and `/?users=alincoln,htaft` (from an external program) are equivalent inputs.
- Rename all the form parameters to be more digestable (`authorPHIDs` -> `authors`). Almost all of them were in this form already anyway. This just gives us `?users=alincoln` instead of `userPHIDs=alincoln`.
- Inside ApplicationSearch, if a request has no query associated with it but does have query parameters, build a query from the request instead of issuing the user's default query. Basically, this means that `/differential/` runs the default query, while `/differential/?users=x` runs a custom query.
Test Plan: {F56612}
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2625, T3775
Differential Revision: https://secure.phabricator.com/D6840
Summary: Fixes T3773. By default, the `/users/` datasource excludes disabled users (since it doesn't make sense to assign them tasks or make them reviewers, for example). However, for ApplicationSearch it does make sense to look for objects, e.g., authored by a disabled user.
Test Plan: Searched for disabled users in Differential.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T3773
Differential Revision: https://secure.phabricator.com/D6834
Summary: Ref T2852. Bleh, gross. Does what it says in the title.
Test Plan: {F54024}
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6735
Summary:
Ref T2852. We need to distinguish between an API call which worked but got back nothing (404) and an API call which failed.
In particular, Asana hit a sync issue which was likely the result of treating a 500 (or some other error) as a 404.
Also clean up a couple small things.
Test Plan: Ran syncs against deleted tasks and saw successful syncs of non-tasks, and simulated random failures and saw them get handled correctly.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2852
Differential Revision: https://secure.phabricator.com/D6470
Summary: Add `getReviewerStatus` to get an array of `DifferentialReviewer` objects. The method `needReviewerStatus` in `DifferentialRevisionQuery` loads the edges into the revisions loaded.
Test Plan: Added `->needReviewerStatus(true)` to `DifferentialRevisionSearchEngine` and checked through logging that the data was being loaded correctly.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D6450
Summary:
Ref T603. Ref T2625. Fixes T3241. Depends on D5451. Depends on D6346.
@wez, this changes the Differential revision list UI substantially and may generate a lot of bikeshedding / who-moved-my-cheese churn. See T3417 for context, for example. The motivations for this change are:
- The list now works on devices, like phones and tablets. This is a requirement to make the rest of Differential work on devices.
- Although ApplicationSearch intentionally presents a simpler interface initially and some options which were one click away before aren't now, it is much more powerful than the search it replaces and allows users to build, save, share, fork, edit, and customize a much wider range of queries. Users who used the old filters frequently can use Advanced Search -> Save Custom Query to create new versions of them, and of any other query. "Edit Queries.." allows users to remove and reorder queries, including builtin queries. Basically, there are like three things which have gone from "1-click" to "a few clicks", and ten trillion things which have gone from "hard/impossible" to "relatively easy".
The local screenshots look a bit iffy, but I think a lot of this is the fakenesss of my test data. If they still feel iffy in production we can tweak them until they feel good, like we did for Maniphest.
Test Plan:
{F48477}
{F48478}
Reviewers: btrahan, chad, wez
Reviewed By: btrahan
CC: aran, s
Maniphest Tasks: T603, T2625, T3241
Differential Revision: https://secure.phabricator.com/D6347
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
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 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 T2222. Currently, we load inline comments by `commentID` here, but we always pass every commentID associated with the revision. Instead, just load non-draft comments by revision ID. This simplifies querying a little bit and is likely faster anyway (draft comments are currently loaded separately).
Test Plan: Looked at some revisions and verified inlines showed up correctly and in the right places.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D6270
Summary: Ref T2222. I didn't translate this query properly; reproduce the original.
Test Plan: When viewing a revision with non-draft inline comments by a user other than the viewer, the inline comments now appear on the changesets themselves.
Reviewers: kawakami, btrahan, garoevans
Reviewed By: garoevans
CC: aran, mbishopim3
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D6281
Summary:
Ref T2222. See D6260.
Push all this junk behind a Query so I can move the storage out from underneath it.
Test Plan: Viewed home page, list view, revision. Made draft, looked at preview, submitted draft, viewed inline, replied to inline.
Reviewers: btrahan
Reviewed By: btrahan
CC: chad, aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D6262
Summary:
Ref T2222.
I'm thinking about how I want to approach the Asana sync, and I want to try to do T2222 first so that we can build it cleanly on top of ApplicationTransactions. I think we can at least walk down this road a little bit and if it turns out to be scary we can take another approach.
I was generally very happy with how the auth migration turned out (seemingly, it was almost completely clean), and want to pursue a similar strategy here. Basically:
- Wrap the new objects in the old objects for reads/writes.
- Migrate all the existing data to the new table.
- Everything hard is done; move things over a piece at a time at a leisurely pace in lots of smallish, relatively-easy-to-understand changes.
This deletes or abstracts all reads of the DifferentialComment table. In particular, these things are **deleted**:
- The script `undo_commits.php`, which I haven't pointed anyone at in a very long time.
- The `differential.getrevisionfeedback` Conduit method, which has been marked deprecated for a year or more.
- The `/stats/` interface in Differential, which should be rebuilt on Fact and has never been exposed in the UI. It does a ton of joins and such which are prohibitively difficult to migrate.
This leaves a small number of reading interfaces, which I replaced with a new `DifferentialCommentQuery`. Some future change will make this actually load transactions and wrap them with DifferentialComment interfaces.
Test Plan: Viewed a revision; made revision comments
Reviewers: btrahan
Reviewed By: btrahan
CC: edward, chad, aran
Maniphest Tasks: T2222
Differential Revision: https://secure.phabricator.com/D6260
Summary:
- Unify all the reference/embed Remarkup rules for Differential, Maniphest, Paste and Ponder.
- Add rules for Pholio.
- Does not yet unify Diffusion or Files (both are a bit more involved).
- Prepare for hovercards.
Test Plan: {F33894}
Reviewers: chad, vrana
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D5120
Summary:
If the revision doesn't have reviewers then it's really not waiting on someone else and the author must take an action.
An improvement would be to check if the reviewers are not disabled but that would require loading their handles.
Test Plan:
/
/differential/
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin, s.o.butler
Differential Revision: https://secure.phabricator.com/D5046
