Summary:
I'm about 90% sure this fixes the intermittent test failure on `testObjectSubscribersPolicyRule()` or whatever.
We use `spl_object_hash()` to identify objects when passing hints about policy changes to policy rules. This is hacky, and I think it's the source of the unit test issue.
Specifically, `spl_object_hash()` is approximately just returning the memory address of the object, and two objects can occasionally use the same memory address (one gets garbage collected; another uses the same memory).
If I replace `spl_object_hash()` with a static value like "zebra", the test failure reproduces.
Instead, sneak an object ID onto a runtime property. This is at least as hacky but shouldn't suffer from the same intermittent failure.
Test Plan: Ran `arc unit --everything`, but I never got a reliable repro of the issue in the first place, so who knows.
Reviewers: chad
Reviewed By: chad
Differential Revision: https://secure.phabricator.com/D17029
Summary: Ref T11957.
Test Plan:
- Viewed an existing project profile.
- Viewed a user profile.
- Created a new project.
- Edited a profile menu.
- Added new profile items.
- Grepped for renamed symbols.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11957
Differential Revision: https://secure.phabricator.com/D17028
Summary:
Ref T11957. This renames the Configuration storage, transaction, query, and PHID type.
No rename on the actual menu item types yet, that's next (and should be the end of this, I think).
Test Plan:
- Viewed projects.
- Viewed profiles.
- Edited a project menu.
- Grepped for all renamed symbols, I think?
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11957
Differential Revision: https://secure.phabricator.com/D17027
Summary: Fixes T11990. Chrome/Ubuntu doesn't respect `outline: auto`, so this is the only workaround I could fine. Somewhat nicer than stock Chrome.
Test Plan: tab tab tab tab, ubuntu, chrome, windows, safari.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Maniphest Tasks: T11990
Differential Revision: https://secure.phabricator.com/D17023
Summary:
- Fixes T11995. This got moved but I missed renaming this callsite.
- Fixes T11993. If you have valid credentials, but haven't run `storage upgrade` yet, we can hit this exception during setup. Just ignore it instead.
Test Plan:
- Saved global settings, no more fatal.
- Changed `storage-namespace` to junk, loaded web UI with valid database credentials.
{F2106358}
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11993, T11995
Differential Revision: https://secure.phabricator.com/D17024
Summary:
- Fixes T11995. This got moved but I missed renaming this callsite.
- Fixes T11993. If you have valid credentials, but haven't run `storage upgrade` yet, we can hit this exception during setup. Just ignore it instead.
Test Plan:
- Saved global settings, no more fatal.
- Changed `storage-namespace` to junk, loaded web UI with valid database credentials.
{F2106358}
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11993, T11995
Differential Revision: https://secure.phabricator.com/D17024
Summary: Allows applications to be added as profile menu items
Test Plan: Add an application to a project, see menu item, click on menu. Uninstall application, see menu without application.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Differential Revision: https://secure.phabricator.com/D17016
Summary: Fixes T11982. If an install is not public, the registering user may not be able to see the inviting user.
Test Plan: {F2097656}
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11982
Differential Revision: https://secure.phabricator.com/D17015
Summary: Fixes T11544. Attempt to detect if we're on a tiny, burstable-CPU AWS instance and complain.
Test Plan:
- Completely faked this locally.
- Hit the URI on an EC2 instance to check that it's correct (got back "m3.large", since that was the instance class).
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11544
Differential Revision: https://secure.phabricator.com/D17014
Summary:
Fixes T11955. Currently, milestones have an internal name of "Parent (Milestone) ...".
This makes them look like they're prefix matches for "Parent", but they're actually prefix matches for "Milestone".
Reorder the names so that the internal name is "Milestone Parent ...".
Test Plan: Created a project "AAA" with milestone "BBB". Searched for "AAA", found "AAA" and milestone "AAA (BBB)".
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11955
Differential Revision: https://secure.phabricator.com/D17013
Summary:
Fixes T11955. Milestone names are currently tokenizing and indexing awkwardly. For example, "A (B C D)" becomes the tokens "A", "(B", "C" and "D)".
The token "(B" can't be searched for since "(" is tokenized on the client.
Instead, tokenize "A (B C D)" into "A", "B", "C", "D".
Test Plan:
- Added unit tests.
- Used `bin/search index --type project --force` to reindex.
- Searched for "A", "B", "C", "D", etc., for real examples.
- Now, found milestones more consistently.
- Also serached for `viewer()`, `members()`, etc.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11955
Differential Revision: https://secure.phabricator.com/D17012
Summary: Fixes T11982. Currently, configuring a custom logo and then setting the policy restrictively locks off the whole install.
Test Plan:
- Configured `ui.logo`.
- Searched for the file PHID in global search to find the underlying file.
- Set the policy to something restrictive ("only me").
- Purged cache (`bin/cache purge --purge-all`).
- Restarted webserver to nuke APC.
- Loaded a page as a different user.
- Before change: policy exception while trying to load the logo.
- After change: fallback to default logo.
- Loaded page as user who can see the logo, got custom logo.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11982
Differential Revision: https://secure.phabricator.com/D17011
Summary:
Ref T11553. With some regularity, users make various configuration mistakes which we can detect by making a request to ourselves.
I use a magical header to make this request because we want to test everything else (parameters, path).
- Fixes T4854, probably. Tries to detect mod_pagespeed by looking for a header. This is a documentation-based "fix", I didn't actually install mod_pagespeed or formally test this.
- Fixes T6866. We now test for parameters (e.g., user somehow lost "QSA").
- Ref T6709. We now test that stuff is decoded exactly once (e.g., user somehow lost "B").
- Fixes T4921. We now test that Authorization survives the request.
- Fixes T2226. Adds a setup check to determine whether gzip is enabled on the web server, and attempts to enable it at the PHP level.
- Fixes `<space space newline newline space><?php` in `preamble.php`.
Test Plan: Tested all of these setup warnings, although mostly by faking them.
Reviewers: joshuaspence, chad
Reviewed By: chad
Subscribers: Korvin
Maniphest Tasks: T4854, T4921, T6709, T6866, T11553, T2226
Differential Revision: https://secure.phabricator.com/D12622
Summary:
Ref T11954. In cluster configurations, we get repository information by making HTTP calls over Conduit.
These are slower than local calls, so clustering imposes a performance penalty. However, we can use futures and parallelize them so that clustering actually improves overall performance.
When not running in clustered mode, this just makes us run stuff inline.
Test Plan:
- Browsed Git, Mercurial and Subversion repositories.
- Locally, saw a 700ms wall time page drop to 200ms.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D17009
Summary:
Fixes T11960. In D16998 I removed some code which validated settings on read to improve performance, but lost this replacement validation in shuffling the patch stack.
This restores similar validation before we write the cache. This has the same effect, it's just faster.
Also, bump the cache key to wipe out anything that got bitten (like my account on `secure` rendering dates wrong).
Test Plan:
- Edited settings, verified the edits held.
- Faked invalid settings, saw the check throw exceptions.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11960
Differential Revision: https://secure.phabricator.com/D17008
Summary: Use one set of spacing everywhere for action menus
Test Plan: Review dropdowns, action lists.
Reviewers: epriestley, tyhtest
Reviewed By: tyhtest
Subscribers: Korvin
Differential Revision: https://secure.phabricator.com/D17006
Summary: Form inputs with just text or certain selects had extra space. This uses better text centering.
Test Plan: Review forms on all Settings pages, Custom Policy, Herald.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Differential Revision: https://secure.phabricator.com/D17005
Summary:
Ref T11954. Especially with higher-latency file stores like S3, we can spend a lot of time reading README data and then pulling it out of file storage.
Instead, cache it.
Test Plan: Browsed a repostory with a README, saw faster pages.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D17002
Summary:
Ref T11954. This is kind of complex and I'm not sure I want to actually land it, but it gives us a fairly good improvement for clustered repositories so I'm leaning toward moving forward.
When we make (or receive) clustered repository requests, we must first load a bunch of stuff out of Almanac to figure out where to send the request (or if we can handle the request ourselves).
This involves several round trip queries into Almanac (service, device, interfaces, bindings, properties) and generally is fairly slow/expensive. The actual data we get out of it is just a list of URIs.
Caching this would be very easy, except that invalidating the cache is difficult, since editing any binding, property, interface, or device may invalidate the cache for indirectly connected services and repositories.
To address this, introduce `PhabricatorCacheEngine`, which is an extensible engine like `PhabricatorDestructionEngine` for propagating cache updates. It has two modes:
- Discover linked objects (that is: find related objects which may need to have caches invalidated).
- Invalidate caches (that is: nuke any caches which need to be nuked).
Both modes are extensible, so third-party code can build repository-dependent caches or whatever. This may be overkill but even if Almanac is the only thing we use it for it feels like a fairly clean solution to the problem.
With `CacheEngine`, make any edit to Almanac stuff propagate up to the Service, and then from the Service to any linked Repositories.
Once we hit repositories, invalidate their caches when Almanac changes.
Test Plan:
- Observed a 20-30ms performance improvement with `ab -n 100`.
- (The main page making Conduit calls also gets a performance improvement, although that's a little trickier to measure directly.)
- Added debugging code to the cache engine stuff to observe the linking and invalidation phases.
- Made invalidation throw; verified that editing properties, bindings, etc, properly invalidates the cache of any indirectly linked repositories.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D17000
Summary:
Ref T11954. This reduces how much work we need to do to load settings, particularly for Conduit (which currently can not benefit directly from the user cache, because it loads the user indirectly via a token).
Specifically:
- Cache builtin defaults in the runtime cache. This means Phabricator may need to be restarted if you change a global setting default, but this is exceptionally rare.
- Cache global defaults in the mutable cache. This means we do less work to load them.
- Avoid loading settings classes if we don't have to.
- If we missed the user cache for settings, try to read it from the cache table before we actually go regenerate it (we miss on Conduit pathways).
Test Plan: Used `ab -n100 ...` to observe a ~6-10ms performance improvement for `user.whoami`.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D16998
Summary:
Ref T11954. I want to store some lists/arrays in the mutable (database) cache, but it only supports string storage.
Provide a serializing wrapper which flattens when values are written and expands them when they're read.
Test Plan: Used by D16997. See that revision.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D16999
Summary:
Ref T11954. When we query for Conduit tokens, we load the associated objects (users) by PHID.
Currently, querying objects by PHID requires us to load every PHIDType class, when we can know which specific classes we actually need (e.g., just `UserPHIDType`, if only user PHIDs are present in the query).
Use PhabricatorCachedClassMapQuery to reduce the number of classes we load on this pathway.
Test Plan:
- Used `ab -n100` to roughly measure a ~5% performance improvement?
- This measurement feels a little flimsy but the XHProf profile is cleaner, at least.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D16997
Summary: Ref T11954. This old query class can use slightly more modern code.
Test Plan: Ran Conduit methods, verified results are unchanged.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D16996
Summary: Ref T11954. Depends on D16994. This implements the Conduit method cache described in that revision for a small global Conduit performance improvement.
Test Plan: Verified Conduit has the same behavior at lower cost. See D16994 for details.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D16995
Summary:
Ref T11954. Depends on D16993. We have a couple of "look up the class for this key" queries which are costly enough to show up on a profile.
These aren't huge wins, but they're pretty easy. We currently do this like this:
```
$class_map = load_every_subclass();
return idx($class_map, $key);
```
However, we don't need to load EVERY subclass if we're only looking for, say, the Conduit method subclass which implements `user.whoami`. This allows us to cache that map and find the right class efficiently.
This cache is self-validating and completely safe even in development.
Test Plan:
- Used `curl` to make queries to `user.whoami`, verified that content was identical before and after the change.
- Used `ab -n100` to roughly measure 99th percentile time, which dropped from 74ms to 65ms. This is a small improvement (13% in the best case, here) but it benefits every Conduit method call.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D16994
Summary:
Ref T11954. Depends on D16992. We have some data which can be generated and cached at runtime. Three examples are:
- Class map from Conduit method names to implementing classes.
- Class map from PHID types to implementing classes.
- The main routing map.
None of these are huge wins but they impose global costs and can be shaved down through caching without introducing an enormous amount of new complexity.
The cost to these maps is that sometimes you'll need to restart your webserver, even in development mode if these caches are active. However, in some cases these changes are very rare, and in other cases we can just leave the cache disabled in development mode without a huge complexity cost.
Specifically, the Conduit/PHID type class maps are self-validating and can not go bad, even in development mode.
The routing map will be able to, but I plan to just disable it in development mode.
This provides a general-purpose pure APC cache stack for storing this data.
Test Plan: See future changes.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11954
Differential Revision: https://secure.phabricator.com/D16993
Summary: Ref T929. When viewing a branch, show a few recent differences from the default branch (usually, "master").
Test Plan: {F2079220}
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T929
Differential Revision: https://secure.phabricator.com/D16991
Summary: Ref T929. We've made some UI updates since D15330.
Test Plan: {F2079125}
Reviewers: avivey, chad
Reviewed By: chad
Maniphest Tasks: T929
Differential Revision: https://secure.phabricator.com/D16990
Summary:
This shows the commits list only (Actual `git diff` will show up at a later date).
The inputs are left as text-fields, to allow the form to accept anything that can be resolved. The form is GET, to allow sharing URIs.
The conduit method response array is compatible with that of `diffusion.historyquery`, to make it easy to build
the "history" table.
The hardest part here was, of course, Naming. I think "from" and "onto" are unconfusing, and I'm fairly confident that the "to merge"
instructions are in sync with the actual content of the page.
Test Plan: Look at several "compare" views, with various values of "from" and "onto".
Reviewers: #blessed_reviewers!, epriestley
Subscribers: caov297, 20after4, Sam2304, reardencode, baileyb, chad, Korvin
Maniphest Tasks: T929
Differential Revision: https://secure.phabricator.com/D15330
Summary:
Persona is going to be decommed November 30th, 2016.
It is highly unlikely that anyone is currently using persona as a real
login method at this point.
Test Plan: tried locally to add auth adapter.
Reviewers: chad, #blessed_reviewers, epriestley
Reviewed By: #blessed_reviewers, epriestley
Subscribers: epriestley
Differential Revision: https://secure.phabricator.com/D16371
Summary:
Fixes T11940. In 2.11.0, Git has made a change so that newly-pushed changes are held in a temporary area until the hook accepts or rejects them.
This magic temporary area is only readable if the appropriate `GIT_ENVIRONMENTAL_MAGIC` variables are available. When executing `git` commands, pass them through from the calling context.
We're intentionally conservative about which variables we pass, and with good reason (see "httpoxy" in T11359). I think this continues to be the correct default behavior.
Test Plan:
- Upgraded to Git 2.11.0.
- Tried to push over SSH, got a hook error.
- Applied patch.
- Pulled and pushed over SSH.
- Pulled and pushed over HTTP.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11940
Differential Revision: https://secure.phabricator.com/D16988
Summary:
Ref T11939. IPv4 addresses can normally only be written in one way, but IPv6 addresses have several formats.
For example, the addresses "FFF::", "FfF::", "fff::", "0ffF::", "0fFf:0::", and "0FfF:0:0:0:0:0:0:0" are all the same address.
Normalize all addresses before writing them to logs, etc, so we store the most-preferred form ("fff::", above).
Test Plan:
Ran an SSH clone over IPv6:
```
$ git fetch ssh://local@::1/diffusion/26/locktopia.git
```
It worked; verified that address read out of `SSH_CLIENT` sensibly.
Faked my remote address as a non-preferred-form IPv6 address using `preamble.php`.
Failed to login, verified that the preferred-form version of the address appeared in the user activity log.
Made IPv6 requests over HTTP:
```
$ curl -H "Host: local.phacility.com" "http://[::1]/"
```
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11939
Differential Revision: https://secure.phabricator.com/D16987
Summary:
Ref T11939. Depends on D16984. Now that CIDRLists can contain IPv6 addresses, blacklist all of the reserved IPv6 space.
This reserved blacklist is used to prevent users from accessing internal services via "Import Calendar" or "Add Macro".
They can't actually reach IPv6 addresses via these mechanisms yet because we need to do more work to support outbound IPv6 requests, but make sure reserved IPv6 space is blacklisted already when that support eventaully arrives.
Also, clean up some error messages (e.g., for trying to hit a bad URI in "Add Macro").
Test Plan:
- Loaded pages with default blacklist.
- Tried to make requests into IPv6 space.
- Currently, this is impossible because of `parse_url()` and `gethostynamel()` calls.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11939
Differential Revision: https://secure.phabricator.com/D16986
Summary: The default behavior of these methods is to return `true`, so these overrides have no effect.
Test Plan: `grep`; poked around.
Reviewers: chad
Reviewed By: chad
Subscribers: hach-que
Differential Revision: https://secure.phabricator.com/D16985
Summary:
Fixes T11946. When a logged-out viewer is loading a page on a non-public install, there are two policy issues which prevent them from loading global settings:
- They can not see the Settings application itself.
- They can not see the global settings object.
Allow them to see Settings by making mandatory applications always visible. (This doesn't make any application pages public.)
Allow them to see the global settings object explicitly.
Test Plan:
Changed default language, viewed logged-out page:
{F2076924}
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11946
Differential Revision: https://secure.phabricator.com/D16983
Summary: Ref T3612. Hides badges on the comment panel.
Test Plan: Give myself a badge, leave a comment, see no badge UI.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Maniphest Tasks: T3612
Differential Revision: https://secure.phabricator.com/D16979
Summary: Ref T3612, prevents lightbox from spawning from inside a lightbox.
Test Plan: Click on file lightbox, leave file comment, click file comment, get take to file page instead of another lightbox.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Maniphest Tasks: T3612
Differential Revision: https://secure.phabricator.com/D16978
Summary: Ref T3612, this adds a anchor around the large icon with hover state so you can download from here as well.
Test Plan: Hover over .ics file, click, get download.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Maniphest Tasks: T3612
Differential Revision: https://secure.phabricator.com/D16977
Summary:
Fixes T11938.
Note that there's a subcase here: if you `hg clone` or `svn checkout` a short `/source/` URI that ends in `.git`, we miss the lookup and don't get this far, so you still get a generic error message.
Hopefully it is clear enough on its own that `proto://.../blah.git` is, in fact, a Git repository, since it says ".git" at the end.
If that doesn't prove to be true, we can be more surgical about this.
Test Plan:
```
$ git clone ssh://local@localvault.phacility.com/source/quack.notgit/
Cloning into 'quack.notgit'...
phabricator-ssh-exec: This repository ("quack.notgit") is not a Git repository. Use "hg" to interact with this repository.
fatal: Could not read from remote repository.
Please make sure you have the correct access rights
and the repository exists.
```
```
$ hg clone ssh://local@localvault.phacility.com/source/phabx
remote: phabricator-ssh-exec: This repository ("phabx") is not a Mercurial repository. Use "git" to interact with this repository.
abort: no suitable response from remote hg!
```
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11938
Differential Revision: https://secure.phabricator.com/D16976
Summary: Ref T11801. In some cases, this could lead to us failing to generate the first recurrence in a series.
Test Plan: Imported `weekly.ics` (from D16974) and saw an event correctly occur on Aug 18, with my local timezone set to "America/Los_Angeles".
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11801
Differential Revision: https://secure.phabricator.com/D16975
Summary:
Fixes T11936. After editing a repository URI, we were not correctly updating the URI index.
Any other edit to the repository //would// update the index, and this index is only really used by `arc` to figure out which repository a working copy belongs to, so that's how this evaded detection for this long. In particular, creating a repository would usually have an edit after any URI edits, to activate it, which would build the index correctly.
Test Plan:
- Added a new URI to a repository.
- Verified it was immediately reflected in the `repository_uriindex` table.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11936
Differential Revision: https://secure.phabricator.com/D16972
Summary:
Ref T11922. When we deploy on Saturday I need to rebuild all the cluster indexes, but some instances won't have anything indexed so they won't actually trigger the activity.
Add a `--force` flag that just clears an activity even if the activity is not required.
Test Plan: Ran `bin/config done reindex --force` several times.
Reviewers: chad
Reviewed By: chad
Maniphest Tasks: T11922
Differential Revision: https://secure.phabricator.com/D16970
Summary: Fixes T11791. We do this in durable column, but not in regular Conpherence. I think this is the right place? Not sure how this will feel with high lag.
Test Plan: Submit lots of text in a Conpherence.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin
Maniphest Tasks: T11791
Differential Revision: https://secure.phabricator.com/D16969
