Summary:
Fixes T4463. When your VCS or account password is not set, we test it for upgrade anyway. This doesn't make sense and throws shortly into the process because the empty hash isn't parseable.
Instead, only show upgrade prompts when the password exists.
Test Plan:
- Added a password to an existing account with no password via password reset.
- Added a VCS password to an existing account with no VCS password.
- Observed no fatals / nonsense behaviors.
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T4463
Differential Revision: https://secure.phabricator.com/D8282
Summary: Fixes T4443. Plug VCS passwords into the shared key stretching. They don't use any real stretching now (I anticipated doing something like T4443 eventually) so we can just migrate them into stretching all at once.
Test Plan:
- Viewed VCS settings.
- Used VCS password after migration.
- Set VCS password.
- Upgraded VCS password by using it.
- Used VCS password some more.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4443
Differential Revision: https://secure.phabricator.com/D8272
Summary:
Ref T4443. In addition to performing upgrades from, e.g., md5 -> bcrypt, also allow sidegrades from, e.g., bcrypt(cost=11) to bcrypt(cost=12). This allows us to, for example, bump the cost function every 18 months and stay on par with Moore's law, on average.
I'm also allowing "upgrades" which technically reduce cost, but this seems like the right thing to do (i.e., generally migrate password storage so it's all uniform, on average).
Test Plan:
- Fiddled the bcrypt cost function and saw appropriate upgrade UI, and upgraded passwords upon password change.
- Passwords still worked.
- Around cost=13 or 14 things start getting noticibly slow, so bcrypt does actually work. Such wow.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4443
Differential Revision: https://secure.phabricator.com/D8271
Summary:
Ref T4443.
- Add a `password_hash()`-based bcrypt hasher if `password_hash()` is available.
- When a user logs in using a password, upgrade their password to the strongest available hash format.
- On the password settings page:
- Warn the user if their password uses any algorithm other than the strongest one.
- Show the algorithm the password uses.
- Show the best available algorithm.
Test Plan: As an md5 user, viewed password settings page and saw a warning. Logged out. Logged in, got upgraded, no more warning. Changed password, verified database rehash. Logged out, logged in.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T4443
Differential Revision: https://secure.phabricator.com/D8270
Summary:
Ref T4443. Make hashing algorithms pluggable and extensible so we can deal with the attendant complexities more easily.
This moves "Iterated MD5" to a modular implementation, and adds a tiny bit of hack-glue so we don't need to migrate the DB in this patch. I'll migrate in the next patch, then add bcrypt.
Test Plan:
- Verified that the same stuff gets stored in the DB (i.e., no functional changes):
- Logged into an old password account.
- Changed password.
- Registered a new account.
- Changed password.
- Switched back to master.
- Logged in / out, changed password.
- Switched back, logged in.
- Ran unit tests (they aren't super extensive, but cover some of the basics).
Reviewers: btrahan
Reviewed By: btrahan
CC: aran, kofalt
Maniphest Tasks: T4443
Differential Revision: https://secure.phabricator.com/D8268
Summary: This allows users to set their HTTP access passwords via Diffusion interface.
Test Plan: Clicked the "Set HTTP Access Password" link, set a password and saw it appear in the DB.
Reviewers: #blessed_reviewers, hach-que, btrahan
Reviewed By: hach-que
CC: Korvin, epriestley, aran, jamesr
Maniphest Tasks: T2230
Differential Revision: https://secure.phabricator.com/D7462
Summary: Ref T3684 for discussion. This could be cleaned up a bit (it would be nice to draw entropy once per request, for instance, and maybe respect CSRF_TOKEN_LENGTH more closely) but should effectively mitigate BREACH.
Test Plan: Submitted forms; submitted forms after mucking with CSRF and observed CSRF error. Verified that source now has "B@..." tokens.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T3684
Differential Revision: https://secure.phabricator.com/D6686
Summary:
Ref T2632. Fixes T1466.
Currently, we normalize slugs (and thus Phriction URIs and canonical project names) to a small number of latin characters. Instead, blacklist a few characters and permit everything else (including utf8 characters).
When generating Phriction URIs, encode any utf8 characters. This means we render URIs encoded, but browsers handle this fine and display them readably in the URI and address bar, etc.
The blacklisted characters are mostly for practical reasons: \x00-\x19 are control characters, `#%?` are meaningful in URIs, `+` is sometimes configured to be interprted as space by apache, etc., `<>\\` are just silly, `&= ` are largely cosmetic.
This allows some silly stuff, like generating URIs with zero-width spaces and RTL markers in them. Possibly we should go blacklist those characters at some point.
Depends on: D5191
Test Plan: {F34402}
Reviewers: AnhNhan, chad, vrana
Reviewed By: chad
CC: aran
Maniphest Tasks: T1466, T2632
Differential Revision: https://secure.phabricator.com/D5192
Summary: Does this seem reasonable? It's a bit more compact than digest() (6 bits / byte instead of 4 bits / byte) and 72 bits of entropy @ 12 bytes instead of 128 bits of entropy @ 32 bytes. I feel like it's important to preserve the printability, though, and this seemed like a fairly good balance of concerns.
Test Plan: unit tests
Reviewers: vrana
Reviewed By: vrana
CC: aran, yemao932
Differential Revision: https://secure.phabricator.com/D4253
Summary:
This commit doesn't change license of any file. It just makes the license implicit (inherited from LICENSE file in the root directory).
We are removing the headers for these reasons:
- It wastes space in editors, less code is visible in editor upon opening a file.
- It brings noise to diff of the first change of any file every year.
- It confuses Git file copy detection when creating small files.
- We don't have an explicit license header in other files (JS, CSS, images, documentation).
- Using license header in every file is not obligatory: http://www.apache.org/dev/apply-license.html#new.
This change is approved by Alma Chao (Lead Open Source and IP Counsel at Facebook).
Test Plan: Verified that the license survived only in LICENSE file and that it didn't modify externals.
Reviewers: epriestley, davidrecordon
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2035
Differential Revision: https://secure.phabricator.com/D3886
Summary:
We currently cache all connections in LiskDAO so we can roll back transactions when fixtured unit tests complete.
Since we establish a new connection wrapper each time we establish a global lock, this cache currently grows without bound.
Instead, pool global lock connections so we never have more than the largest number of locks we've held open at once (in PullLocalDaemon, always 1).
Another way to solve this is probably to add an "onclose" callback to `AphrontDatabaseConnection` so that it can notify any caches that it been closed. However, we currently allow a connection to be later reopened (which seeems reasonable) so we'd need a callback for that too. This is much simpler, and this use case is unusual, so I'd like to wait for more use cases before pursing a more complicated fix.
Test Plan:
Ran this in a loop:
while (true) {
for ($ii = 0; $ii < 100; $ii++) {
$lock = PhabricatorGlobalLock::newLock('derp');
$lock->lock();
$lock->unlock();
}
$this->sleep(1);
}
Previously it leaked ~100KB/sec, now has stable memory usage.
Reviewers: vrana, nh, btrahan
Reviewed By: vrana
CC: aran
Maniphest Tasks: T1636
Differential Revision: https://secure.phabricator.com/D3239
Summary: Add an explicit close() method to connections and call it in GlobalLock.
Test Plan:
Wrote a script like this:
$lock = PhabricatorGlobalLock::newLock('test');
echo "LOCK";
$lock->lock();
sleep(10);
echo "UNLOCK";
$lock->unlock();
sleep(9999);
Using `SHOW FULL PROCESSLIST`, verified the connection closed after 10 seconds with both the "MySQL" and "MySQLi" implementations.
Reviewers: btrahan, vrana
Reviewed By: vrana
CC: aran
Maniphest Tasks: T1470
Differential Revision: https://secure.phabricator.com/D3035
Summary:
Currently, multiple unit tests that acquire global locks will interfere with each other. Namespace the locks so they don't.
(Possibly we should also rename this to PhabricatorStorageNamespaceLock or something since it's not really global any more, but that's kind of unwieldy...)
Test Plan: Acquired locks with --trace and verified they were namespaced properly.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1162
Differential Revision: https://secure.phabricator.com/D2939
Summary: See D2924.
Test Plan: Ran locks with blocking timeouts.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D2925
Summary: Implementation is a little crazy but this seems to work as advertised.
Test Plan: Acquired locks with "lock.php". Verified they held as long as the process reamined open and released properly on kill -9, ^C, etc.
Reviewers: nh, jungejason, vrana, btrahan, Girish, edward
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1400
Differential Revision: https://secure.phabricator.com/D2864
Summary:
- `kill_init.php` said "Moving 1000 files" - I hope that this is not some limit in `FileFinder`.
- [src/infrastructure/celerity] `git mv utils.php map.php; git mv api/utils.php api.php`
- Comment `phutil_libraries` in `.arcconfig` and run `arc liberate`.
NOTE: `arc diff` timed out so I'm pushing it without review.
Test Plan:
/D1234
Browsed around, especially in `applications/repository/worker/commitchangeparser` and `applications/` in general.
Auditors: epriestley
Maniphest Tasks: T1103
Summary:
- For context, see T547. This is the last (maybe?) in a series of diffs that
moves us off raw sha1() calls in order to make it easier to audit the codebase
for correct use of hash functions.
- This breaks CSRF tokens. Any open forms will generate an error when
submitted, so maybe upgrade off-peak.
- We now generate HMAC mail keys but accept MAC or HMAC. In a few months, we
can remove the MAC version.
- The only remaining callsite is Conduit. We can't use HMAC since Arcanist
would need to know the key. {T550} provides a better solution to this, anyway.
Test Plan:
- Verified CSRF tokens generate properly.
- Manually changed CSRF to an incorrect value and got an error.
- Verified mail generates with a new mail hash.
- Verified Phabricator accepts both old and new mail hashes.
- Verified Phabricator rejects bad mail hashes.
- Checked user log, things look OK.
Reviewers: btrahan, jungejason, benmathews
Reviewed By: btrahan
CC: aran, epriestley, btrahan
Maniphest Tasks: T547
Differential Revision: 1237
