Summary:
this is a request from Facebook:
> someone has added macros for common words like 'push', 'lgtm'.
> We want to let he image macro page at least attribute to an owner so
> that we can publicly shame whoever added the 800px 'clowntown' macro.
Test Plan:
test macros with/without author (all macro should have
author. This is just to be safe).
Reviewers: epriestley, nh
Reviewed By: epriestley
CC: hwang, aran, arice, jungejason, epriestley
Differential Revision: 984
Summary: There are currently two files, but all scripts require both of them,
which is clearly silly. In the longer term I want to rewrite all of this init
stuff to be more structured (e.g., merge webroot/index.php and __init_script__
better) but this reduces the surface area of the ad-hoc "include files" API we
have now, at least.
Test Plan:
- Grepped for __init_env__.php (no hits)
- Ran a unit test (to test unit changes)
- Ran a daemon (to test daemon changes)
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 976
Summary:
This is a pretty bad, but working implmentation of a way to kill one particular PID that
is controlled by Phabricator. Also remove the second 'stop' from the ##phd help## list.
Test Plan:
[ricky@rhelpad01 phabricator] (phd-stop-twice)$ ./bin/phd status
PID Started Daemon
30154 Oct 1 2011, 2:38:08 AM PhabricatorMetaMTADaemon
30172 Oct 1 2011, 2:38:09 AM PhabricatorMetaMTADaemon
30190 Oct 1 2011, 2:38:09 AM PhabricatorMetaMTADaemon
30210 Oct 1 2011, 2:38:09 AM PhabricatorMetaMTADaemon
[ricky@rhelpad01 phabricator] (phd-stop-twice)$ ./bin/phd stop 30190
Stopping daemon 'PhabricatorMetaMTADaemon' (30190)...
Daemon 30190 exited normally.
[ricky@rhelpad01 phabricator] (phd-stop-twice)$ ./bin/phd stop 123456
123456 is not controlled by Phabricator. Not killing.
[ricky@rhelpad01 phabricator] (phd-stop-twice)$ ./bin/phd stop
Stopping daemon 'PhabricatorMetaMTADaemon' (30154)...
Stopping daemon 'PhabricatorMetaMTADaemon' (30172)...
Stopping daemon 'PhabricatorMetaMTADaemon' (30210)...
Daemon 30210 exited normally.
Daemon 30154 exited normally.
Daemon 30172 exited normally.
Reviewers: epriestley
CC:
Differential Revision: 975
Summary: Shows events which a page dispatched, plus all the registered
listeners.
Test Plan:
Pretty basic for now, but works OK:
https://secure.phabricator.com/file/view/PHID-FILE-49fcd23081ce55cf9369/
(I also made it dispatch some dummy events to verify they show up.)
Reviewers: aran
Reviewed By: aran
CC: aran
Differential Revision: 973
Summary:
This is an attempt to satisfy a lot of the one-off requests a little more
generally, by providing a relatively generic piece of event architecture.
Allow the registation of event listeners which can react to various application
events (currently, task editing).
I'll doc this a bit better but I wanted to see if anyone had massive objections
to doing this or the broad approach. The specific problem I want to address is
that one client wants to do a bunch of routing for tasks via email, so it's
either build a hook, or have them override most of ManiphestReplyHandler, or
something slightly more general like this.
Test Plan: Wrote a silly listener that adds "Quack!" to a task every time it is
edited and edited some tasks. I was justly rewarded.
Reviewers: nh, jungejason, tuomaspelkonen, aran
Reviewed By: aran
CC: aran, epriestley
Differential Revision: 881
Summary: See T507. Since you can't do "xxxxxxxx^" where "xxxxxxxx" is the first
commit in a repository, fall back to diffing against the empty tree if we fail
to diff against the parent commit.
Test Plan: Looked at the first commit in libphutil on my local.
Reviewers: edward, jungejason, nh, tuomaspelkonen, aran
Reviewed By: nh
CC: aran, edward, epriestley, nh
Differential Revision: 953
Summary: The display of images pairs is not corresponding to the selected two
image diffs. The fix is to use reference to get the phid for each image.
Test Plan: Create a revision with two diffs of images.
Test the display between base and diff1/diff2.
Test the rendering of images between diff1 and diff2.
Test the inline comments also.
Reviewers: epriestley, jungejason
CC:
Differential Revision: 955
Summary:
We need to issue all commands as $repository->junk() so we can pick up
credentials. Some of this stuff predates that change landing.
(I removed the "https" vs "svn+ssh" fallback code since it's specific to
Facebook, affected a tiny number of commits, is basically an SVN bug with UTF-8
handling and HTTP support, and doesn't make sense in the general case. The user
has the tools they need to force it via "reparse.php" if it's really an issue.)
Test Plan: Created new authenticated-remote mercurial and git repositories and
pulled/discovered them with credentials.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 970
Summary: Change import script plus almost all the view stuff. Still some rough
edges but this seems to mostly work. Blame is currently unsupported but I think
everything else works properly.
Test Plan:
Imported the hg repository itself. It doesn't immediately seem completely
broken. Here are some screens:
https://secure.phabricator.com/file/view/PHID-FILE-1438b71cc7c4a2eb4569/https://secure.phabricator.com/file/view/PHID-FILE-3cec4f72f39e7de2d041/https://secure.phabricator.com/file/view/PHID-FILE-2ea4883f160e8e5098f9/https://secure.phabricator.com/file/view/PHID-FILE-35f751a36ebf65399ade/
All the parsers were able to churn through it without errors.
Ran the new "reparse.php" script in various one-commit and repository modes.
Browsed/imported some git repos for good measure.
NOTE: The hg repository is only 15,000 commits and around 1,000 files.
Performance is okay but hg doesn't provide performant, native APIs to get some
data efficiently so we have to do some dumb stuff. If some of these interfaces
are cripplingly slow or whatever, let me know and we can start bundling some
Mercurial extensions with Arcanist.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde, epriestley
Differential Revision: 960
Summary: See D943, this is the second parse stage. This will mark Differential revisions as "Committed" among other things.
Almost all the logic here is shared between VCSes so the implementation itself is straightforward.
Test Plan: Parsed all messages for the official Mercurial repository.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
CC:
Differential Revision: 944
Summary:
Splitting up D960 a bit, see that for context.
We currently have two scripts, "parse_one_commit.php" and
"reparse_all_commit_messages.php", but they're sort of silly and you can't do
certain things with them. Replace them with one script which is more flexible
and can do specific reparse steps on individual commits or entire repos.
I left the old scripts as stubs since I think there are some FB wiki docs and
stuff that mention them. I'll delete them in a month or whenever I remember or
something.
Test Plan: Ran "reparse.php" with various arguments, including vs-one-commit,
vs-repository, with --trace, and against different types of repos.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 964
Summary:
Postponed unit tests are not unit tests with problems. The results
just haven't arrived yet.
Test Plan: Tested accepting a diff with unit status 1, 3, 5 (ok, errors,
postponed)
Reviewers: jungejason, epriestley
Reviewed By: epriestley
CC: aran, epriestley, tuomaspelkonen
Differential Revision: 969
Summary:
Just breaking D960 into some smaller parts, this is a standalone method used in
Mercurial parsing.
(There's a bad version of this function in the SVN stuff but I'll get rid of it
the next time I'm in there.)
Test Plan: See D960.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 965
Summary:
When we discover a new commit and it has a known local commit or tree hash, mark
it committed.
This supports Mercurial and Git-Immutable workflows, and improves
hybrid-Git-Mutable workflows and covers some cases where poeple just make
mistakes or whatever.
Test Plan: Parsed Mercurial, Git and SVN commits.
Reviewers: Makinde
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 963
Summary: See T502. Under some VCS setups, we get full email addresses instead of
usernames or real names. Try harder to find matches, by falling back to email
address parsing if we don't get hits on the straight-up token parsing.
Test Plan:
This is difficult to test because it depends on the account state and repository
state, and hard to pull out so it's more testable without better mocking
facilities. I just dumped this into the parser to verify the behavior:
foreach (array(
'epriestley',
'Evan Priestley',
'epriestley@epriestley.com',
'derp <epriestley@epriestley.com>',
'"Evan Priestley" <derpderpderp@derpderpderp.com>',
'quackderp <derpderpderp@derpderpderp.com>',
) as $email) {
echo "{$email} = ".$this->resolveUserPHID($email)."\n";
}
die();
Running PhabricatorRepositoryGitCommitMessageParserWorker...
epriestley = PHID-USER-79f25616ea2635089a31
Evan Priestley = PHID-USER-79f25616ea2635089a31
epriestley@epriestley.com = PHID-USER-1bec59b91be6223f07fd
derp <epriestley@epriestley.com> = PHID-USER-1bec59b91be6223f07fd
"Evan Priestley" <derpderpderp@derpderpderp.com> =
PHID-USER-79f25616ea2635089a31
quackderp <derpderpderp@derpderpderp.com> =
This is expected (all variations of my identity parsed correctly, and the bogus
one failed). There are two different user PHIDs in the result set because I have
like 30 different similar accounts on my local, including one called "derp" and
another one with address "derp@derp.com", which prevented an earlier version of
this test case from working correctly.
Reviewers: zachallia, aran, Makinde, jungejason, nh, tuomaspelkonen
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 968
Summary:
- We throw on a missing date right now, in the DateTime constructor. This can
happen in reasonable cases and this is display code, so handle it more
gracefully (see T520).
- This stuff is a little slow and we sometimes render many hundreds of dates
per page. I've been seeing it in profiles on and off. Memoize timezones to
improve performance.
- Some minor code duplication that would have become less-minor with the
constructor change, consolidate the logic.
- Add some unit tests and a little documentation.
Test Plan:
- Ran unit tests.
- Profiled 1,000 calls to phabricator_datetime(), cost dropped from ~49ms to
~19ms with addition of memoization. This is still slower than I'd like but I
don't think there's an easy way to squeeze it down further.
Reviewers: ajtrichards, jungejason, nh, tuomaspelkonen, aran
Reviewed By: ajtrichards
CC: aran, ajtrichards, epriestley
Differential Revision: 966
datasources
Summary:
The open source Phabricator has like 3,500 user accounts now and it takes a
while to pull/render them. Add an option to switch to ondemand for large
installs.
I'll follow up with a patch at some point to address a couple of name things:
- Denormalize last names into a keyed column (although this evidences some
bias toward the western world).
- Force all usernames to lowercase (sorry Girish, Makinde).
Also this patch is so clean it's crazy.
Didn't bother with other object types for now, I'm planning to dedicate a few
days to Projects at some point and I'll flesh out some auxiliary features like
this when I do that.
Test Plan: Switched to ondemand, verified data was queried dynamically. Switched
back, verified data was preloaded.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: nh
CC: aran, epriestley, nh
Differential Revision: 923
Summary:
We currently rely on "remote_hooks_enabled" in .arcconfig to determine whether
commands like "arc amend" and "arc merge" should imply "arc mark-committed".
However, this is a historical artifact that is now bad for a bunch of reasons:
- The option name is confusing, it really means 'repository is tracked'.
- The option is hard to discover and generally sucks.
- We can empirically determine the right answer since we now know if a project
is in a tracked repository.
Add a call which arcanist can make on these workflows to figure out if it is
interacting with a project in a tracked repository or not.
Also added an "isTracked()" convenience method to reduce the number of magic
strings all over the place.
Test Plan: Ran "arcanist.projectinfo" for nonexistent, untracked and tracked
projects.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, epriestley, Makinde
Differential Revision: 945
Summary:
See D943, this is the second parse stage. This will mark Differential revisions
as "Committed" among other things.
Almost all the logic here is shared between VCSes so the implementation itself
is straightforward.
Test Plan: Parsed all messages for the official Mercurial repository.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 944
Summary:
Repository import has three major steps:
- Commit discovery (serial)
- Message parsing (parallel, mostly VCS independent)
- Change parsing (parallel, highly VCS dependent)
This implements commit discovery for Mercurial, similar to git's parsing:
- List the heads of all the branches.
- If we haven't already discovered them, follow them back to their roots (or
the first commit we have discovered).
- Import all the newly discovered commits, oldest first.
This is a little complicated but it ensures we discover commits in depth order,
so the discovery process is robust against interruption/failure. If we just
inserted commits as we went, we might read the tip, insert it, and then crash.
When we ran again, we'd think we had already discovered commits older than HEAD.
This also allows later stages to rely on being able to find Phabricator commit
IDs which correspond to parent commits.
NOTE: This importer is fairly slow because "hg" has a large startup time
(compare "hg --version" to "git --version" and "svn --version"; on my machine,
hg has 60ms of overhead for any command) and we need to run many commands (see
the whole "hg id" mess). You can expect something like 10,000 per hour, which
means you may need to run overnight to discover a large repository (IIRC, the
svn/git discovery processes are both about an order of magnitude faster). We
could improve this with batching, but I want to keep it as simple as possible
for now.
Test Plan: Discovered all the commits in the main Mercurial repository,
http://selenic.com/repo/hg.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 943
Summary: See D939. Regardless of what we do there, these will break, and they're
pretty silly anyway (see the giant caveat comments in the second one).
Test Plan: Clicked a direct-jump comment link, did save/cancel for inline
comments.
Reviewers: phil, cpojer, tomo, mroch
Reviewed By: phil
CC: aran, phil
Differential Revision: 940
Summary: Explains how to use the immutable history doctrine and mercurial.
Recommends "one idea is one commit".
Test Plan: Read documentation.
Reviewers: fratrik, Makinde, aran, jungejason, tuomaspelkonen, cpiro
Reviewed By: cpiro
CC: aran, cpiro, epriestley, ide
Differential Revision: 861
Summary:
dirname('x') returns '.', not '/'; this caused some issues for repositories with
files at the root.
There are some cases in the parsers where I should probably swap this out too
but I'll wait until I'm doing some more rigorous testing since that stuff is a
bit fragile and this fixes an immediate issue.
Test Plan: Ran unit tests. Viewed a file at root level in a test repository.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: nh
CC: aran, nh
Differential Revision: 932
Summary: See T262. This creates the index on the Differential side which we need in order to execute this query efficiently on the Diffusion side.
Also renames "DiffusionGitPathIDQuery" to "DiffusionPathIDQuery", this query object has nothing to do with git.
Test Plan: Attached top-level and sub-level diffs to revisions and verified they populated the table with sensible data.
Reviewers: bmaurer, aravindn, fmoo, jungejason, nh, tuomaspelkonen, aran
CC:
Differential Revision: 931
Summary:
we use to only add X-Frame-Options for AphrontWebpageResponse.
There some security concern about it. Example of a drag-drop attack:
http://sites.google.com/site/tentacoloviola/. The fix is to add it to
all AphrontResponse.
Test Plan:
View page which disalble this option still works (like the
xhpast tree page); verify that the AphrontAjaxResponse contains the
X-Frame-Options in the header.
Reviewers: epriestley, benmathews
Reviewed By: epriestley
CC: nh, aran, jungejason, epriestley
Differential Revision: 926
Summary: Feedback from @makinde. These are easy (and necessary) to configure so
we might as well give the user a heads up.
Test Plan: Regenerated the documentation and read "Configuration Guide".
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 929
Summary: For reasons explained in the config I've omitted this from the default
action set, but it's trivial to support it. See D916.
Test Plan: Commented on a revision, was informed I could "!accept" in the email.
Used "!accept" to accept the revision.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 928
Summary: This method relies on 64-bit math being available, which isn't a safe
assumption. Use the builtin bc functions instead for arbitrarily large integers.
Test Plan: @skrul, can you apply this locally and let me know if it works?
Reviewers: skrul, hunterbridges, jungejason, nh, tuomaspelkonen, aran
Reviewed By: skrul
CC: aran, skrul, epriestley
Differential Revision: 912
Summary:
This is pretty straightforward, except:
- We need to request read/write access to the address book to get the account
ID (which we MUST have) and real name, email and account name (which we'd like
to have). This is way more access than we should need, but there's apparently no
"get_loggedin_user_basic_information" type of call in the Google API suite (or,
at least, I couldn't find one).
- We can't get the profile picture or profile URI since there's no Plus API
access and Google users don't have meaningful public pages otherwise.
- Google doesn't save the fact that you've authorized the app, so every time
you want to login you need to reaffirm that you want to give us silly amounts of
access. Phabricator sessions are pretty long-duration though so this shouldn't
be a major issue.
Test Plan:
- Registered, logged out, and logged in with Google.
- Registered, logged out, and logged in with Facebook / Github to make sure I
didn't break anything.
- Linked / unlinked Google accounts.
Reviewers: Makinde, jungejason, nh, tuomaspelkonen, aran
Reviewed By: aran
CC: aran, epriestley, Makinde
Differential Revision: 916
Summary: No actual parsing/import yet, but now you can define and pull Mercurial
repositories. I merged most of the local pull code so we can share it between
hg/git.
Test Plan:
- Created a new Mercurial repository to track Codeigniter off Bitbucket
- Edited / saved / etc.
- Launched the mercurial pull daemon, it pulled the repo. Killed and
relaunched, it updated the repo.
- Launched the git fetch deamon, it still works correctly.
Reviewers: Makinde, aran, jungejason, tuomaspelkonen
Reviewed By: Makinde
CC: aran, Makinde
Differential Revision: 793
Summary:
Changed the documentation to describe the project-subproject join table
instead of the task-project join table.
Test Plan:
none
Reviewers:
epriestley, cadamo
CC:
Differential Revision: 927
Summary:
The CSRF changes meant that we can't generate a file URI with just its PHID
anymore, and converted a mathematical function into a service call.
Unfortunately, this caused massive perf problems in some parts of the
application, critically handles, where loading N users became N single gets.
Derp derp derp. Remedy this by doing a single multiget. This substantially
improves performance of many interfaces, particularly the Maniphest task list.
I need to go through the rest of the PhabricatorFileURI callsites and get rid of
them, but I think this is the most substantive one.
Test Plan: Profiled Maniphest task list, queries went from >100 to a handful.
Explosion of multiderp. :/ Looked at some views with profile photos to verify
they still render accurately.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: aran
CC: aran
Differential Revision: 921
Summary: This will get fancier, but here's a basic interface for doing symbol
lookups. Still all pretty tentative.
Test Plan: Looked up various things, got some sensible results.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: tuomaspelkonen
CC: aran, tuomaspelkonen
Differential Revision: 900
Summary: See T315 for an extensive description of this feature. Adds the
descibed storage table.
Test Plan: Used phpsh to read/write symbol objects.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: tuomaspelkonen
CC: aran, epriestley, tuomaspelkonen
Differential Revision: 897
Summary: This got caught in the CSRF filter but is a safe write.
Test Plan: Pasted the URI for a picture of a goat into a diff, saw a goat.
Reviewers: aran, jungejason
Reviewed By: aran
CC: aran
Differential Revision: 910
Summary: See D902. As @abdul notes, a password input is probably more
appropraite here.
Test Plan: Mashed stuff into it, got bullets instead of text.
Reviewers: abdul, jungejason, nh, tuomaspelkonen, aran
Reviewed By: nh
CC: aran, nh
Differential Revision: 913
Summary: phriction.edit allows you to omit the content string, meaning "don't
edit content". If you're also creating the page, we currently break in a
terrible, horrible, no-good, very-bad way because 'content' can't be null.
Default to empty string instead so phriction.edit creates an empty page instead
of a broken mess.
Test Plan: Called phriction.edit on a new page with no content.
Reviewers: skrul, jungejason, nh, tuomaspelkonen, aran
Reviewed By: skrul
CC: aran, skrul
Differential Revision: 920
Summary:
When Diffusion encounters an image file, it displays it as an
image, but when it encounters a PDF file, it currently shows only some
gibberish. This fixes that.
Test Plan:
I tried it. Embedding a large PDF in a data URL is a little
bit slow, but it works.
Reviewers: tuomaspelkonen, epriestley, gc3, waltermundt, jungejason, nh
Reviewed By: epriestley
CC: aran, tuomaspelkonen, epriestley, jaapweel
Differential Revision: 915
Summary:
We need to query the unit status in order to determine if
there are postponed unit tests to update after running "arc unit"
Test Plan:
1) set my conduit uri to a server running the new code
2) ensured unitStatus existed when retrieving a diff
Reviewers: epriestley
Reviewed By: epriestley
CC: dpepper, aran, epriestley
Differential Revision: 918
Summary:
@tomo ran into an issue where he had some non-SSL-only cookie or whatever, so
"Logout" had no apparent effect. Make sure "Logout" really works by destroying
the session.
I originally kept the sessions around to be able to debug session stuff, but we
have a fairly good session log now and no reprorted session bugs except for all
the cookie stuff. It's also slightly more secure to actually destroy sessions,
since it means "logout" breaks any cookies that attackers somehow stole (e.g.,
by reading your requests off a public wifi network).
Test Plan: Commented out the cookie clear and logged out. I was logged out and
given a useful error message about clearing my cookies.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: aran
CC: tomo, aran, epriestley
Differential Revision: 911
- Removed irrelevant csprintf(..)
- Updated code to use $repository->getRemoteURI()
- Updated code to use getRemoteCommandFuture(..) in Diffusion code
- Updated code to use $repository->getRemoteURI()
Summary: I still need to go through all the daemon and Diffusion code and change
the bare execx() calls to $repository->execxXXX() to actually make this work,
but we're getting close.
Test Plan: Configured repositories with various HTTP / SVN setups and ran the
test_connection.php script to verify keys were located and added and
username/password information was supplied.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: nh
CC: aran, nh, jungejason
Differential Revision: 902
setup
Summary: See T481. We'll fail the pcntl test if we don't have this, in a
potentially confusing way. Test and detect missing 'php' explicitly before we
try the pcntl test, so we can give the user a better error message.
Test Plan: In setup mode, did a good run and then faked it to execute 'phpx'
instead to get a failure.
Reviewers: johnduhart, jungejason, tuomaspelkonen, aran
Reviewed By: tuomaspelkonen
CC: aran, epriestley, tuomaspelkonen
Differential Revision: 878
Summary: Added line number 1 for each image and added code to display the
comments for each image.
Test Plan: Adding an image in my local directory and create a revision for it.
Click line number 1, and the comment window prompts. Adding and save the
comment. The comment shows in the differential comment list and in the inline
comment. Submit the comment. Create more comments for the image and the
"Previous" and "Next" buttons all work well.
Reviewers: epriestley, jungejason
CC:
Differential Revision: 901
Summary: Previously, this code accidentally did not use the best URI. Instead,
use the best URI. It's the best, obviously.
Test Plan: Uploaded a binary file and then clicked the preview.
Reviewers: hunterbridges, jungejason, nh, tuomaspelkonen, aran
Reviewed By: tuomaspelkonen
CC: aran, tuomaspelkonen
Differential Revision: 905
Summary:
Move toward storing credentials in configuration so it's easier to get the
daemons working. This should eventually solve all the key juggling junk you have
to do right now.
This only gets us part of the way to actually using these credentials in the
daemons since I have to go swap everything for $repository->execBlah().
I tried to write a web "Test Connection" button but it was too much of a mess to
get git to work since git doesn't give you access to its SSH command and SSH has
a bunch of interactive prompts which you can't really do anything about without
it or a bunch of ~/.ssh/config editing. This is what Git recommends:
https://git.wiki.kernel.org/index.php/GitFaq#How_do_I_specify_what_ssh_key_git_should_use.3F
..but it's not a great match for this use case.
Test Plan:
- Only partial.
- Ran "test_connection.php" on a Git repo with and without SSH, and with and
without valid credentials. This part works properly.
- Ran "test_connection.php" on a public SVN repo, but I don't have private or
WEBDAV repos set up at the moment.
- Mercurial doesn't work yet.
- Daemons haven't been converted yet.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, abdul, nmalcolm, epriestley, jungejason
Differential Revision: 888
Summary:
These fields use auxiliary storage now. Migrate the data and get rid of the
columns in the main table.
- This might take a little while to run, although there are <500k rows so
probably not too long.
- Maybe grab a backup of the table first, if I screwed something up this will
delete the data in these fields.
Test Plan:
- Ran migration locally.
- Browsed Differential.
- Grepped for "revertPlan" and "blameRevision".
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason, epriestley
Differential Revision: 832
Summary:
Give users better errors and UI:
- For subpath SVN repositories, default the path to the subdirectory, not to
"/". This makes the home screen useful and things generally less confusing.
- For unparsed commits, show a more descriptive error message without the
"blah blah" silliness.
- For paths outside of the subpath parse tree, short circuit into an
appropriate error message.
- For foreign SVN stub commits (see D892), show an explicit message.
Test Plan: Looked at unparsed commits, subpath repositories, foreign stub
commits, and paths outside of the subpath parse tree. Received sensible error
messages.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 894
Summary: These queries are safe to run without a CSRF token, and we need them
for the query analyzer in DarkConsole.
Test Plan: "Analyze Query Plans" works again.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: nh
CC: aran, epriestley, nh
Differential Revision: 895
Summary:
In password-based auth environments, there is now a user settings
panel to allow them to change their password.
Test Plan:
Click settings, choose password from the left:
* enter current password, new password (twice), log out, and log in with
new password
* enter current password, non-matching passwords, and get error
* enter invalid old password, and get error
* use firebug to change csrf token and verify that it does not save with
and invalid token
Changed config to disable password auth, loaded settings panel and saw
that password was no longer visible. Tried loading the panel anyway and
got redirected.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: 890
references
Summary:
See T325. We tentatively support doing partial subdirectory parses in
Phabricator for Subversion, so you can elect to import only "trunk/local/" or
similar. We do this by importing only some of the commits (those commits which
affected that directory).
In Subversion, you can also "svn cp
svn+ssh://example.com/svnroot/trunk/foreign/example.c@13 local.c". This means
that commits which reference "trunk/local/" may themselves reference foreign
commits.
Currently, we break in this case and can't find the commit reference. Instead,
generate a foreign commit stub so we can at least point at some reasonable
object.
Test Plan: Successfully imported trunk/a/ of the test repo in T325 without
errors. Verified commit 3 in that repo is imported as a foreign stub.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason, epriestley
Differential Revision: 892
repositories
Summary:
This query isn't scoped correctly to the repository ID, so we may identify
commits from other repositories.
This causes a somewhat subtle issue since we only use it to manage file
copies/moves, so you end up with a file "copied from" the same revision in
another repository. I think the UI probably even renders correctly.
Once I finish T325 and better understand what's going on here, I'll see how much
work is involved in writing an SQL patch to fix this.
Test Plan: Parsed the test repo from T325 with the expected error.
Reviewers: jungejason, nh, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 891
Summary:
The differential panels at the top of the differential revision view page
were 2px smaller than the divs on the bottom of the page (everything below
the table of contents). This diff makes differential-panel 2px wider so it
matches.
Test Plan: viewed a differential revision and checked that the divs lined up
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: 887
Summary: See T489. Provide slightly more detail so we can figure out if there's
a real issue here.
Test Plan:
Hit URIs like:
/differential/comment/preview/29/
/differential/comment/preview/29/?__ajax__=1
/differential/comment/preview/29/?__csrf__=1
..and got appropriate error messages.
Reviewers: jungejason
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 884
Summary:
oh god everyone hates this
revert revert
https://www.facebook.com/photo.php?fbid=787360256660&set=p.787360256660&type=1&theater
(I left the icons themselves since I have some plans to do other things with
them.)
Test Plan: I am not good at designer
Reviewers: ola, elynde, bh, ashwin, jungejason, kdelong, zrait, tomo, aran
Reviewed By: aran
CC: aran, epriestley, tomo
Differential Revision: 885
Summary:
I didn't realize createDiffDict was a public method when I
modified it, and I broke the API call in getrevision. This moves the
modification inside the method and reverts the method header back to
it's original form.
Test Plan: none
Reviewers: epriestley
Reviewed By: epriestley
CC: edward, aran, epriestley
Differential Revision: 883
Summary:
After D857, we try to attach local commit information to revisions. If this
information is available, display it on the revision.
Design on this is a little rough, I might try to combine this into the revision
update view or something like that since we're starting to take up a lot of real
estate for metadata.
Test Plan: Local diffed this and got some commit info.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, epriestley, jungejason
Differential Revision: 872
Summary: Allow Conduit method so they stop raising lint warnings. See D874.
Test Plan: Ran "arc lint" on conduit files and was no longer given frivolous
warnings.
Reviewers: nh, jungejason, tuomaspelkonen, aran
Reviewed By: nh
CC: aran, nh, epriestley
Differential Revision: 875
Summary: If the link text is a URI, just treat it as a nameable (and possibly
relative) URI link. See tasks.
Test Plan: Copy/pasted the doc example into Phriction, links worked.
Reviewers: skrul, hunterbridges, jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 882
don't show "Undo"
Summary: When a user hits "Reply" on an inline comment, doesn't type anything,
and then hits "Cancel", we incorrectly store the text of the comment the user is
replying to as the "original" text, and then detect that they've changed it when
they immediately cancel. Instead, store empty string as the original text.
Test Plan:
- Hit "Reply" and then "Cancel" on an inline comment. No undo now.
- Hit "Reply", typed some text, and then hit "Cancel". Got an undo which
restored my text.
Reviewers: tomo, jungejason, tuomaspelkonen, aran
Reviewed By: aran
CC: aran, tomo
Differential Revision: 879
Summary: HPHP has behaviorial differences from PHP which make this logic
problematic and we provide a good error message to users when there's a cookie
issue now, so unsplit the cookie logic and just clear the same cookie we'd
otherwise set, as per ssl / base domain.
Test Plan: Logged in and out of my local install.
Reviewers: jungejason
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 876
Summary:
Provide three Phriction methods for programmatic access to Phriction:
- phriction.info: get information about a document
- phriction.history: get change history for a document
- phriction.edit: create or update a document
I moved all the editing/creation logic into a new PhrictionDocumentEditor to
share code between the Conduit and Web edit pathways.
Test Plan: Got info and history via conduit. Edited and created new pages via
conduit and web.
Reviewers: hsb, jungejason, tuomaspelkonen, aran, hunterbridges
Reviewed By: hunterbridges
CC: skrul, aran, hunterbridges
Differential Revision: 866
Summary:
We need createlintresults because we are doing extended
static analysis offline, and thus we need to be able to update the
lint results associated with a diff. This is similar to
updateunitresults, but "create" is more accurate than "update" since
we never need to change existing lint results.
getdiffproperty is used by the client to ensure it isn't creating any
duplicates lint results. It's the symmetric operation to
setdiffproperty, which already exists.
Test Plan:
We have a new offline linter that I used to test. This
linter calls getdiffproperty on every run.
1. Tested updating an existing set of lint results by first running
"arc diff" with lint errors caught by the local linter, then later
running offline analysis which catches one other error and updates via
createlintresults. Ensured the differential lint results were as
expected.
2. Tested the creation of an entirely new diff property through
createlintresults. I first ran "arc diff --nolint" to skip all lint
results, then ran offline analysis which caught an error and updated
through createlintresults. Ensured differential lint results were as
expected.
Reviewers: epriestley
Reviewed By: epriestley
CC: dpepper, aran, mgummelt, jungejason, epriestley
Differential Revision: 868
Summary: Oops, I left this in from an earlier version and missed it since I was
mostly looking at Maniphest for testing. We already render this information in
the header, don't additionally render it under the comments.
Test Plan: derp derp, loaded any revision with sourced comments
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 871
Summary: Execute searches like the primary Maniphest task list. Let me know what
else you guys need from this API.
Test Plan: Executed "maniphest.info" and "maniphest.find"
Reviewers: jungejason, tuomaspelkonen, aran, nh
Reviewed By: nh
CC: blair, skrul, aran, jungejason, epriestley, nh, tuomaspelkonen
Differential Revision: 867
Summary:
When an object is updated, record the content source for the update. This mostly
isn't terribly useful but one concrete thing I want to do with it is let admins
audit via-email replies more easily since there are a bunch of options which let
you do hyjinx if you intentionally configure them insecurely. I think having a
little more auditability around this feature is generally good. At some point
I'm going to turn this into a link admins can click to see details.
It also allows us to see how frequently different mechanisms are used, and lets
you see if someone is at their desk or on a mobile or whatever, at least
indirectly.
The "tablet" and "mobile" sources are currently unused but I figured I'd throw
them in anyway. SMS support should definitely happen at some point.
Not 100% sure about the design for this, I might change it to plain text at some
point.
Test Plan: Updated objects and saw update sources rendered.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, epriestley, jungejason
Differential Revision: 844
Summary: This may make it easier to debug problems with CLI + pcntl
Test Plan: Added a syntax error to the script and got more useful output
Reviewers: johnduhart, jungejason, tuomaspelkonen, aran
Reviewed By: johnduhart
CC: aran, johnduhart
Differential Revision: 869
Summary: When a user stores the empty string in an auxiliary field, simply don't
store it, and delete it if it already exists.
Test Plan: Edited a revision with an empty "Quack" field, got an empty row in
the DB. Applied patch, edited empty again, row went away. Edited empty again,
still no row. Edited and put something in the field, got a row.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 865
Summary: "set -e" causes the script to exit if any command returns nonzero.
Without it, we incorrectly discard the failure code. At Facebook everything runs
in "set -e" or some equivalent so I never picked this up in testing.
Test Plan:
Added the hook to my svn local, it blocked bad commits and allowed good ones.
>>> orbital:~/devtools/svnroot $ svn commit -m "quaa"
Sending test.php
Transmitting file data .svn: Commit failed (details follow):
svn: Commit blocked by pre-commit hook (exit code 1) with output:
LINT ERRORS
This changeset has lint errors. You must fix all lint errors before you can
commit.
You can add '@bypass-lint' to your commit message to disable lint checks for
this commit, or '@nolint' to the file with errors to disable lint for that
file.
>>> Lint for test.php:
Error (XHP1) PHP Syntax Error!
This file contains a syntax error: XHPAST Parse Error: syntax error,
unexpected '}' on line 1
>>> 1 <?php asdddddd;!}}
Reviewers: svemir, jungejason, tuomaspelkonen, aran
Reviewed By: svemir
CC: aran, svemir, epriestley
Differential Revision: 864
Summary: cpiro reported a cache inconsistency issue from a push a while ago
which this should fix (see #?????), and we haven't sync'd in a while anyway.
Test Plan: Poked some interfaces very gently.
Reviewers: cpiro, cpojer, tomo, jungejason, tuomaspelkonen, aran
Reviewed By: tomo
CC: aran, epriestley, tomo, cpiro
Differential Revision: 859
Summary:
It is possible to view a comment that has no cache; when viewing such a comment
the request doesn't have a csrf token and there is no need for one, so we turn
off the write guard.
Test Plan:
loaded an old diff that had no cache, and the page loaded instead of throwing
an AphrontCSRFException.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: 858
Summary: Quora wants to handle some moderation tasks with Phabricator, but want
to lower the barrier to entry for the install and let moderators adopt it
gradually. One request is to allow auth rules to be relaxed so we can auth based
on Reply-To to make things easier. This is insecure if configured but not really
a big deal and the patch isn't big or complicated.
Test Plan: Sent a test email with bogus "From" but valid "Reply-To". It was
rejected with this setting off, and allowed with this setting on.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 842
Summary: Some stack frames do not have file/line information, e.g. __autoload
triggers. Render these as "Internal".
Test Plan: Reloaded a trace with an internal __autoload() frame, got
"(Internal)" instead of ": 0" with warnings.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 843
Summary: Do JOIN in the protocol handler, after we receive 376 ("end of motd").
Test Plan: Ran bot, it joined a channel after receieving a 376 command.
Reviewers: moos3, codeblock, aran, jungejason, tuomaspelkonen
Reviewed By: moos3
CC: aran, moos3
Differential Revision: 855
Summary: Delete one line which has no effect.
Test Plan: Open revision page to make sure it still works.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: 852
Summary: Instead of just saying a task is "Closed", say "Resolved", "Wontfix",
etc.
Test Plan: Looked at task list view, saw "Resolved", "Wontfix", etc.
Reviewers: skrul, hunterbridges, jungejason, tuomaspelkonen, aran
Reviewed By: skrul
CC: aran, skrul, jungejason, epriestley
Differential Revision: 851
Summary: If a user partially discovers a repository and then deletes it, the
timeline will have events from the old repository which this daemon won't be
able to parse.
Test Plan: @ajtrichards, can you apply this locally and restart your daemons
(##phd stop##, then relaunch them) and let me know if it fixes the issue?
Reviewers: ajtrichards, jungejason, tuomaspelkonen, aran
Reviewed By: ajtrichards
CC: aran, epriestley, ajtrichards
Differential Revision: 845
Summary:
When selecting children of a directory, it is possible that none of its
children exist anymore even though the directory still exists. After fetching
the children but before returning them, we should check whether there are any,
and if there are no children, set the reason as empty directory.
Test Plan:
In sandbox, browsed in diffusion to a directory that exists but has no
files and saw that it has a useful message instead of a vague exception.
Reviewers: epriestley, tuomaspelkonen, jungejason
Reviewed By: tuomaspelkonen
CC: aran, tuomaspelkonen
Differential Revision: 846
"user":"authenticationusername",
"pass":"thisuserspassowrd",
This will allow people with internal irc servers to use this if they control access from ldap for irc.
Summary:
This allows you to configure a single mailbox for all mail sent by phabricator,
so you
can keep a mailaddress like bugs@example.com and don't need a catchall on your
domain/subdomain.
Test Plan:
Enabled and disabled suffix. Saw mails generated have to correct prefix. Also
piped raw mails
into the scripts/mail/mail_handler.php and ensured comments went into
phabricator for both maniphest
and differential.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: 815
Summary: This could be a lot fancier but let's see what else we need. Also fixed
some bugs with maniphest.info.
Test Plan: Used the Conduit web console to create some tasks with different
values.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason, epriestley
Differential Revision: 824
purposes
Summary:
Browsers send port numbers (like ":443" or proxy ports) in the Host header and
we'll currently reject them with a message like:
> Blah is configured on "x.y.com" but you are accessing it on "x.y.com:443".
Instead, examine only the host part.
Test Plan: Had my local listen on port 81 and accessed Phabricator before/after
the change; it now works without throwing.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, epriestley, abdul, jungejason
Differential Revision: 841
Summary: Open AphrontWriteGuard for user login.
Test Plan: verified that the user can log in.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, epriestley
Differential Revision: 840
Summary: When a user has bad cookies, try to clear everything and tell them they
might need to manually clear things.
Test Plan: Added "&& false" to the valid branch and got the exception message.
Reviewers: jungejason
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 839
Summary:
After D814 and D829, you should be able to implement this logic in the
didWriteRevision() method of the field.
Note that the attacher is still referenced in
ConduitAPI_differential_updatetaskrevisionassoc_Method. This method should
probably be moved to facebook/ since it's pretty Facebook-specific.
No rush on any of this, it's not hurting anything.
Test Plan:
- Hit differential.getcommitmessage
- Ran 'arc diff'
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 830
Summary:
In D758, I tightened the scope for which we issue cookies. Instead of setting
them on the whole domain we set them only on the subdomain, and we set them as
HTTPS only if the install is HTTPS.
However, this can leave the user with a stale HTTP cookie which the browser
sends and which never gets cleared. Handle this situation by:
- Clear all four <domain, https> pairs when clearing cookies ("nuke it from
orbit").
- Clear 'phsid' cookies when they're invalid.
Test Plan: Applied a hackier version of this patch to secure.phabricator.com and
was able to login with a stale HTTP cookie.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 838
Summary:
This commit lets users who are filing Maniphest tasks to attache files to them
right off the bat.
Test Plan:
{F3545}
and
{F3546}
Reviewers: epriestley, fmoo, aran
CC:
Differential Revision: 837
Summary:
Remove the blame revision, revert plan and lines fields from the default field
loadout. (After D829 this doesn't cause issues where we have bogus dictionary
entries.)
You should add these back to the Facebook configuration since Facebook wants
these fields. However, I want to keep the default stack very light and I never
saw a huge amount of value in these fields at Facebook so I don't think they
make the cut. Sorry, tomo. ;_;
Test Plan: Ran "arc diff" locally.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: aran
CC: aran, tomo, epriestley
Differential Revision: 831
Differential comments
Summary: If you @mention several users, at least one of which is already CC'd,
we unset all the CCs and don't attach the "Added CCs: ..." block to the comment.
Test Plan: @mentioned two users, one of whom was already CC'd.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason
Differential Revision: 827
Summary:
I think this is the last major step -- use the fields to parse commit messages,
not a hard-coded list of stuff. This adds two primary methods to fields, one to
get all the labels they'll parse (so we can do "CC" and "CCs" and treat them as
the same field) and one to parse the string into a canonical representation
(e.g., lookup reviewers and such).
You'll need to impelement the one block of task-specific stuff I removed in
Facebook's task field:
list($pre_comment) = split(' -- ', $data);
$data = array_filter(preg_split('/[^\d]+/', $pre_comment));
foreach ($data as $k => $v) {
$data[$k] = (int)$v;
}
$data = array_unique($data);
break;
Otherwise I think this is clean.
Test Plan:
- Called the conduit method with various commit messages, parsed fields/errors
seemed correct.
- "arc diff"'d this diff onto localhost, then updated it.
- "arc amend"'d this diff.
Reviewers: jungejason, tuomaspelkonen, aran
Reviewed By: jungejason
CC: aran, jungejason, epriestley
Differential Revision: 829
Summary: While I thought this was complicated, there was nothing subtle or
tricky here -- I just misnamed a variable.
Test Plan: Created a revision with default CCs, got CCs instead of nothing.
Reviewers: aran, jungejason, tuomaspelkonen
Reviewed By: aran
CC: aran
Differential Revision: 834
Summary:
deprecate generateProperties() from class
DifferentialRevisionDetailRenderer. Custom fields now provides a much
more powerful version of generateProperties().
Depends on D814.
Test Plan:
implemented facebook task field with custom field and
verified it worked.
Reviewers: epriestley, tuomaspelkonen
Reviewed By: epriestley
CC: aran, jungejason, epriestley
Differential Revision: 826
Summary:
When rendering commit messages, drive all the logic through field specification
classes instead of the hard-coded DifferentialCommitMessageData class. This
removes DifferentialCommitMessageData and support classes.
Note that this effectively reverts D546, and will cause a minor break for
Facebook (Task IDs will no longer render in commit messages generated by "arc
amend", and will not be editable via "arc diff --edit"). This can be resolved by
implementing the feature as a custom field. While I've been able to preserve the
task ID functionality elsewhere, I felt this implementation was too complex to
reasonably leave hooks for, and the break is pretty minor.
Test Plan:
- Made numerous calls to differential.getcommitmessage across many diffs in
various states, with and without 'edit' and with and without various field
overrides.
- General behavior seems correct (messages look accurate, and have the
expected information). Special fields like "Reviewed By" and "git-svn-id" seem
to work correctly.
- Edit behavior seems correct (edit mode shows all editable fields, hides
fields like "Reviewed By").
- Field overwrite behavior seems correct (overwritable fields show the correct
values when overwritten, ignore provided values otherwise).
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, jungejason
Differential Revision: 814
Summary:
Make the unhandled exception dialogs slightly more useful:
- Make them easier to read.
- Link to files from Phabricator libraries.
- Don't show traces by default.
- Show traces in development mode.
- Rename button from "Cancel" to "Close" and only show it for Ajax.
Test Plan: Rigged DirectoryHomeController to throw, loaded home page. Changed
stack trace setting in config. Clicked some files in the trace.
Reviewed By: aran
Reviewers: aran, jungejason, tuomaspelkonen, codeblock
CC: aran, epriestley
Differential Revision: 823
Summary:
See T354. List every rule which has ever been applied in X-Herald-Rules, not
just the ones which most recently triggered.
Also some random fixes while I was debugging this:
- When conduit methods throw non-conduit exceptions, make sure they get
logged.
- Trigger the Facebook "tasks" backcompat block only if we were going to fail
(this should reduce the shakniess of the transition).
- Fix some log spew from the new field stuff.
Test Plan:
- Created a rule (ID #3) "No Zebras" which triggers for revisions without
"zebra" in the title.
- Created a revision without "zebra" in the title, got X-Herald-Rules: <2>,
<3>
- Updated revision to have "zebra" in the title, verified rule did not trigger
in Herald transcript.
- Verified X-Herald-Rules is still: <2>, <3>
Reviewed By: aran
Reviewers: aran, jungejason, tuomaspelkonen
CC: aran, epriestley
Differential Revision: 817
Summary:
Provide a catchall mechanism to find unprotected writes.
- Depends on D758.
- Similar to WriteOnHTTPGet stuff from Facebook's stack.
- Since we have a small number of storage mechanisms and highly structured
read/write pathways, we can explicitly answer the question "is this page
performing a write?".
- Never allow writes without CSRF checks.
- This will probably break some things. That's fine: they're CSRF
vulnerabilities or weird edge cases that we can fix. But don't push to Facebook
for a few days unless you're prepared to deal with this.
- **>>> MEGADERP: All Conduit write APIs are currently vulnerable to CSRF!
<<<**
Test Plan:
- Ran some scripts that perform writes (scripts/search indexers), no issues.
- Performed normal CSRF submits.
- Added writes to an un-CSRF'd page, got an exception.
- Executed conduit methods.
- Did login/logout (this works because the logged-out user validates the
logged-out csrf "token").
- Did OAuth login.
- Did OAuth registration.
Reviewers: pedram, andrewjcg, erling, jungejason, tuomaspelkonen, aran,
codeblock
Commenters: pedram
CC: aran, epriestley, pedram
Differential Revision: 777
domain
Summary:
See D758, D759.
- Provide a strongly recommended setting which permits configuration of an
alternate domain.
- Lock cookies down better: set them on the exact domain, and use SSL-only if
the configuration is HTTPS.
- Prevent Phabriator from setting cookies on other domains.
This assumes D759 will land, it is not effective without that change.
Test Plan:
- Attempted to login from a different domain and was rejected.
- Logged out, logged back in normally.
- Put install in setup mode and verified it revealed a warning.
- Configured an alterate domain.
- Tried to view an image with an old URI, got a 400.
- Went to /files/ and verified links rendered to the alternate domain.
- Viewed an alternate domain file.
- Tried to view an alternate domain file without the secret key, got a 404.
Reviewers: andrewjcg, erling, aran, tuomaspelkonen, jungejason, codeblock
CC: aran
Differential Revision: 760
Summary: This prevents <applet /> attacks unless the attacker can upload an
applet which has a viewable MIME type as detected by `file`. I'm not sure if
this is possible or not. It should, at least, narrow the attack window. There
are no real tradeoffs here, this is probably a strictly better application
behavior regardless of the security issues.
Test Plan:
- Tried to download a file via GET, got redirected to info.
- Downloaded a file via POST + CSRF from the info page.
Reviewers: andrewjcg, erling, aran, jungejason, tuomaspelkonen
CC: aran
Differential Revision: 759
Summary:
We don't currently validate CSRF tokens on this workflow. This allows an
attacker to upload arbitrary files on the user's behalf. Although I believe the
tight list of servable mime-types means that's more or less the end of the
attack, this is still a vulnerability.
In the long term, the right solution is probably to pass CSRF tokens on all Ajax
requests in an HTTP header (or just a GET param) or something like that.
However, this endpoint is unique and this is the quickest and most direct way to
close the hole.
Test Plan:
- Drop-uploaded files to Files, Maniphest, Phriction and Differential.
- Modified CSRF vaidator to use __csrf__.'x' and verified uploads and form
submissions don't work.
Reviewers: andrewjcg, aran, jungejason, tuomaspelkonen, erling
Commenters: andrewjcg, pedram
CC: aran, epriestley, andrewjcg, pedram
Differential Revision: 758
Summary: See T429. When you hit certain errors, you get less-than-helpful
messages like "upload error 3". Instead, produce human-readable errors.
Test Plan: Simulated errors, verified user receives decent error messages.
Reviewed By: aran
Reviewers: jungejason, tuomaspelkonen, aran, startupguy
CC: aran
Differential Revision: 816
Summary:
When we create or update a revision, we use a parsed commit message dictionary
to edit its fields. Drive consumption of the dictionary through custom fields
instead of hardcoding.
This requires adding some fields which don't really do anything right now to
cover fields which appear only in the commit message.
Test Plan: "arc diff"'d this revision against localhost, "arc diff"'d again to
update.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, jungejason
Differential Revision: 811
Summary:
Move all the rest of the fields into the custom field schema, for revision
views.
I left a couple of stubs in here (willWriteRevision, didWriteRevision) since I'd
planned to do edits here too, but this diff is sort of big-ish already. I'll do
all the edit fields in the next revision.
Depends on D808.
Test Plan: Viewed, edited and conduit'ed some revisions.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, jungejason, epriestley
Differential Revision: 809
Summary:
Move additional fields (which rely on loading handles) to the extensible field
classes and out of hardcoding in the controller.
Depends on D807.
Test Plan: Viewed, edited, and hit conduit for revisions.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, jungejason, epriestley
Differential Revision: 808
Summary:
Differential has a bunch of display-only fields, implement them all as field
specifications instead of hard-coded fields.
Also add some more documentation and fix redundant string constants in blame
rev/revert plan fields.
Test Plan: Viewed, edited, and hit conduit for revisions.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, jungejason, epriestley
Differential Revision: 807
Summary:
- Fix a bug where 'caption' didn't do anything.
- Provide an abstract base implementation for extensions.
- Add some documentation.
- Expose aux fields via conduit.
Test Plan: Added some fields like "Dinosaur", "Kilograms" and "derp" on my local
install. Read documentation.
Reviewed By: jungejason
Reviewers: hunterbridges, jungejason, tuomaspelkonen, aran
CC: aran, philc, jungejason
Differential Revision: 785
Summary: Similar to D785 for Maniphest, expose auxiliary field values via
Conduit.
Test Plan: Ran revision.getinfo on a revision with aux fields, got them in the
response.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, jungejason
Differential Revision: 802
Summary:
This is just to ease transitions for any installs which use these fields (e.g.,
Facebook). I'll write some docs and a migration script once this stuff is a
little more solid, too.
Depends on D800.
Technically these are "better" than the current fields since they show up other
places than the edit screen (derp derp).
Test Plan: Created a field selector which provides these; verified they work by
typing stuff into them and saving the revision.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, epriestley, jungejason
Differential Revision: 801
Summary: Depends on D798. Extends custom fields and makes the vaguely useful:
they can appear on the edit and view interfaces. This does not integrate them
with commit messages yet; that's more complicated but I plan to do it shortly.
Test Plan: Implemented a custom field per P123, it correctly appears on the edit
interface, persists, validates, and shows up when viewing the revision.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, epriestley, jungejason
Differential Revision: 800
Summary:
Precursor to building this out to solve T343. This is similar to the Maniphest
fields we landed recently, although I think they're dissimilar enough that it
isn't worth going crazy trying to make them share code, at least for now.
This doesn't really do anything yet, just adds a storage object and a couple of
selector/field indirection classes.
Test Plan: Ran SQL upgrade script, created an aux field.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran
CC: aran, jungejason
Differential Revision: 798
Summary: Simplify the division of a parent task into several subtasks by looping
the "create subtask" workflow. This replaces "Create Another Task" with "Create
Another Subtask" when you arrive via subtasking.
Test Plan:
- Created a task, looped task create flow.
- Created a subtask, looped subtask create flow.
Reviewed By: codeblock
Reviewers: hunterbridges, codeblock, jungejason, tuomaspelkonen, aran
CC: aran, codeblock, epriestley
Differential Revision: 779
Summary:
we need to know the revision id in order to generate
differential links. It would be amazing if there existed some library
for object <-> network call mapping, and we could get all the
information about an object given some unique ID rather than having to
create a dictionary manually. One can dream...
Test Plan:
1. setup on test phabricator instance
2. called differential.getdiff from client code with a valid diffid
3. verified that the correct revisionID was included
Reviewed By: epriestley
Reviewers: epriestley
CC: dpepper, aran, epriestley, mgummelt
Differential Revision: 795
Summary:
In preparation for adding another search engine (see T355):
- Rename "executor" to "engine".
- Move all engine-specific operations into the engine. Specifically, this
means that indexing moves out of the document store and into the engine (it was
sort of silly where it was before).
- Split choice of an engine into an overridable "selector" class, a base API,
and a concrete MySQL implementation (just like storage engine selection).
- Make all callers go through the indirection layer.
The default selector just unconditionally selects the MySQL engine, but now
(with D786) I can build an Elastic Search engine and you guys can build a
multi-target engine if you want and I don't get there fast enough.
Test Plan:
- Created a new document (task).
- Searched for and found it.
- Viewed index reconstruction.
Reviewed By: jungejason
Reviewers: jungejason, amckinley, tuomaspelkonen, aran
CC: aran, jungejason, epriestley
Differential Revision: 788
Summary:
This is a very small step toward building a Status and possibly an Oncall tool.
Build a calendar view which renders months.
Much of my hesitance to bang these tools out is that dealing with
dates/calendaring is basically horrible, so I'm trying to ease into it.
This calendar is locale-aware and all that jazz.
Test Plan:
- See:
https://secure.phabricator.com/file/view/PHID-FILE-c07a9c663a7d040d2529/
- Verified that months have the right number of days, today is the right day
of the week, months begin on the day after previous months end on, etc.
Reviewed By: aran
Reviewers: jungejason, tuomaspelkonen, aran
Commenters: cwbeck, jungejason
CC: blair, aran, epriestley, cwbeck, jungejason
Differential Revision: 791
Summary:
Provide a {T123} syntax which pulls in the entire name of an object, not just a
link to it. A major use for this is organizing projects using wiki pages. Since
handle links show object status now, this lets you organize stuff in an ad-hoc
way and get a reasonable overview of it. We can make handles richer in the
future, too.
The performance on this isn't perfect (it adds some potential single gets) but I
think it's okay for now and I don't want to make remarkup engine even more
complex until the preprocess/postprocess stuff has had a chance to settle and
I'm more confident it works.
In Differential and Maniphest we'll also incorrectly cache the object
state/name, but that'll fix itself once I move the cache code to use
preprocess/postprocess correctly.
Test Plan:
- See https://secure.phabricator.com/file/view/PHID-FILE-5f9ca32407bec20899b9/
for an example.
- Generated and looked over the documentation.
Reviewed By: jungejason
Reviewers: jungejason, tuomaspelkonen, aran, hunterbridges
CC: skrul, aran, jungejason, epriestley
Differential Revision: 784
