Summary:
Ref T3684. The URI itself is reflected in a few places. It is generally not dangerous because we only let you add random stuff to the end of it for one or two controllers (e.g., the file download controller lets you add "/whatever.jpg"), but:
- Remove it entirely in the main request, since it serves no purpose.
- Remove query parameters in Ajax requests. These are available in DarkConsole proper.
Also mask a few things in the "Request" tab; I've never used these fields when debugging or during support, and they leak quasi-sensitive information that could get screenshotted or over-the-shoulder'd.
I didn't mitgate `__metablock__` because I think the threat is so close to 0 that it's not worthwhile.
Test Plan: Used Darkconsole, examined Requests tab.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T3684
Differential Revision: https://secure.phabricator.com/D6699
Summary: Fixes T3673. Supposedly we won't get any data in this case, but it seems we sometimes do. See discussion in task.
Test Plan: Used `var_dump()`, etc., to verify we short circuit out of "multipart/form-data" posts regardless of the presence of input data.
Reviewers: nmalcolm, btrahan
Reviewed By: nmalcolm
CC: aran
Maniphest Tasks: T3673
Differential Revision: https://secure.phabricator.com/D6670
Summary: The `phabricator.allowed-uris` config setting is not checked properly when trying to set cookies.
Test Plan:
Set an alternate URI, then accessed Phabricator. No longer received a secondary cookie error.
Hit the new exceptions to test them:
{F51131}
{F51132}
Reviewers: btrahan, garoevans
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D6528
Summary: Fixes T3544. Depends on D6475. This was just a missing dependency combined with some questionable error handling which I'll maybe fix some day.
Test Plan: Loaded page, saw result.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T3544
Differential Revision: https://secure.phabricator.com/D6476
Summary: Mostly straightforward. Also fixed a couple of error/darkconsole things.
Test Plan:
- Created poll;
- viewed poll;
- voted in poll;
- used `V6` and `{V6}` markup styles in poll.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Differential Revision: https://secure.phabricator.com/D6458
Summary:
D6278 kind of got closed and commited, this is the actual direction.
Ref T3432
Depends on D6277
Test Plan: Keep using the site
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin, mbishopim3
Maniphest Tasks: T3432
Differential Revision: https://secure.phabricator.com/D6283
Summary:
If D6277 is the way to go, then this will be it's implementation.
Depends on D6277
Test Plan: Keep using the site
Reviewers: epriestley
CC: aran, Korvin
Maniphest Tasks: T3432
Differential Revision: https://secure.phabricator.com/D6278
Summary:
Ref T1536. This is extremely reachable and changes the login code to the new stuff.
Notes:
- I've hard-disabled password registration since I want installs to explicitly flip it on via config if they want it. New installs will get it by default in the future, but old installs shouldn't have their auth options change.
- Google doesn't let us change the redirect URI, so keep the old one working.
- We need to keep a bit of LDAP around for now for LDAP import.
- **Facebook:** This causes substantive changes in what login code is executed.
Test Plan:
- Logged in / logged out / registered, hit new flows.
- Logged in with google.
- Verified no password registration by default.
Reviewers: btrahan, chad
Reviewed By: chad
CC: wez, nh, aran, mbishopim3
Maniphest Tasks: T1536
Differential Revision: https://secure.phabricator.com/D6222
Summary: Ref T1536. Error state is a bit gross but we need to sort that out in general.
Test Plan:
{F46549}
{F46550}
Reviewers: chad, btrahan
Reviewed By: chad
CC: aran
Maniphest Tasks: T1536
Differential Revision: https://secure.phabricator.com/D6208
Summary:
Ref T1445. Ref T1536. Although we have separate CSRF protection and have never been vulnerable to OAuth hijacking, properly implementing the "state" parameter provides a little more certainty.
Before OAuth, we set a random value on the client, and pass its hash as the "state" parameter. Upon return, validate that (a) the user has a nonempty "phcid" cookie and (b) the OAuth endpoint passed back the correct state (the hash of that cookie).
Test Plan: Logged in with all OAuth providers, which all apparently support `state`.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran, arice
Maniphest Tasks: T1445, T1536
Differential Revision: https://secure.phabricator.com/D6179
Summary: Ref T1536. We can safely replace the old login validation controller with this new one, and reduce code dplication while we're at it.
Test Plan: Logged in with LDAP, logged in with OAuth, logged in with username/password, did a password reset.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T1536
Differential Revision: https://secure.phabricator.com/D6178
Summary:
Provide a Content-Length header so that browsers can estimate time
remaining for file downloads.
Test Plan: Tested on our local phabricator install.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D6107
Summary:
Ref T2625. Lifts almost all of the search logic out of Paste controllers and into Search.
This uses controller delegation for generalization. We use this in a few places, but don't use it very much yet. I think it's pretty reasonable as-is, but I might be able to make even more stuff free.
There are some slightly rough edges around routes, still, but I want to hit Phame and Differential (which both have multiple application search engines) before trying to generalize that.
Test Plan: Executed, browsed and managed Paste searches.
Reviewers: btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2625
Differential Revision: https://secure.phabricator.com/D6073
Summary:
Kind of a quick look at an idea for T2184
Ref T2184
Test Plan: Make sure the site still loads
Reviewers: epriestley
CC: aran, Korvin, mbishopim3
Maniphest Tasks: T2184
Differential Revision: https://secure.phabricator.com/D6045
Summary:
Ref T2785
Looks for hosts in `conduit.servers` config and if any exist route any conduit calls through any one of the hosts.
Test Plan:
Make some curl calls to public methods (`conduit.ping`), watch the access log for two requests. Make some calls from the UI that require authentication, watch the access log a bit more.
Also ran the unit tests.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2785
Differential Revision: https://secure.phabricator.com/D5970
Summary:
I always put a `phlog()` somewhere or something fails and I have hard times figuring out which request it was.
Also fix safe HTML in panel.
Test Plan: Looked at DarkConsole with error on main page, AJAX request and both.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D5784
Summary:
Ref T2787. For payment methods that allow you to add a billable method (i.e., a credit card), move all the logic into the provider. In particular:
- Providers may (Stripe, Balanced) or may not (Paypal, MtGox) allow you to add rebillable payment methods. Providers which don't allow rebillable methods will appear at checkout instead and we'll just invoice you every month if you don't use a rebillable method.
- Providers which permit creation of rebillable methods handle their own data entry, since this will be per-provider.
- "Add Payment Method" now prompts you to choose a provider. This is super ugly and barely-usable for the moment. When there's only one choice, we'll auto-select it in the future.
Test Plan: Added new Stripe payment methods; hit all the Stripe errors.
Reviewers: btrahan, chad
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2787
Differential Revision: https://secure.phabricator.com/D5756
Summary:
Refs T1048; Depends on D5542, D5543, D5544 - It currently just renders multiple hovercards nicely for test purposes. More is on the way.
Mode `test`: Human test chamber.
Mode `retrieve`: For JS. Added so it would not clash with search key routing.
badassery
Test Plan:
`/search/hovercard/test/?phids[hover-T4]=PHID-TASK-g5pduvwrrwvkq5gkx736&phids[hover-T2]=PHID-TASK-gta6lzaaagziavkktima`
Verified the appearance of two tasks with correct rendering and correct ids
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T1048
Differential Revision: https://secure.phabricator.com/D5545
Summary:
For some time, we've stopped the profiler twice when it was invoked by the sampling mechanism. The first time it actually stops, and we write a profile. The second time it hadn't been started, so it returns empty and we write an invalid profile.
Instead, keep track of whether it is running or not, and don't stop it a second time.
Ref T2870.
Test Plan: Set sample rate to 1-in-3, observed valid sample profiles generate.
Reviewers: btrahan, chad
CC: aran
Maniphest Tasks: T2870
Differential Revision: https://secure.phabricator.com/D5534
Summary:
We have a fair number of conditionals on the existence of the access log. Instead, always build it and just don't write it if the user doesn't want a version on disk.
Also, formalize logged-in user PHID (avoids object existence juggling) in the access log and move microseconds-since-startup to PhabricatorStartup (simplifies index.php).
Depends on D5532. Fixes T2860. Ref T2870.
Test Plan: Disabled access log, verified XHProf writes occurred correctly.
Reviewers: btrahan, chad
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2860, T2870
Differential Revision: https://secure.phabricator.com/D5533
Summary:
Ref T2870. This resolves a few issues:
- No proper Application. Define one.
- Routes are in the default controller. Move them to the application.
- UI doesn't work on mobile.
- Overescaping in the link column.
Test Plan:
Old page:
{F38444}
New page:
{F38445}
Reviewers: btrahan, chad
Reviewed By: chad
CC: aran, AnhNhan, edward
Maniphest Tasks: T2870
Differential Revision: https://secure.phabricator.com/D5531
Summary: Modern conduit responses should never have a JSON shield. We disable it for normal responses, but uncaught exceptions hit this higher-level handler block which fails to disable the shield. Disable the shield.
Test Plan: Inspection.
Reviewers: btrahan, andrewjcg
Reviewed By: andrewjcg
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D5483
Summary:
Ref T2787. This does very little so far, but makes inroads on accounts and billing. This is mostly just modeled on what Stripe looks like. The objects are:
- **Account**: Has one or more authorized users, who can make manage the account. An example might be "Phacility", and the three of us would be able to manage it. A user may be associated with more than one account (e.g., a corporate account and a personal account) but the UI tries to simplify the common case of a single account.
- **Payment Method**: Something we can get sweet sweet money from; for now, a credit card registered with Stripe. Payment methods are associated with an account.
- **Product**: A good (one time charge) or service (recurring charge). This might be "t-shirt" or "enterprise plan" or "hourly support" or whatever else.
- **Purchase**: Represents a user purchasing a Product for an Account, using a Payment Method. e.g., you bought a shirt, or started a plan, or purchased support.
- **Charge**: Actual charges against payment methods. A Purchase can create more than one charge if it's a plan, or if the first charge fails and we re-bill.
This doesn't fully account for stuff like coupons/discounts yet but they should fit into the model without any issues.
This only implements `Account`, and that only partially.
Test Plan: {F37531}
Reviewers: chad, btrahan
Reviewed By: btrahan
CC: aran
Maniphest Tasks: T2787
Differential Revision: https://secure.phabricator.com/D5435
Summary:
Safari has a weird bug with `border-radius` plus border color:
{F35865}
Move the uncolored borders to an internal div to fix this. Also tweak some positioning on icons for cards, and add a "magenta" color.
Test Plan: {F35866}
Reviewers: chad, btrahan
Reviewed By: chad
CC: aran
Differential Revision: https://secure.phabricator.com/D5338
Summary:
Ref T2650. Possible fix for that issue.
- "Passthru" got renamed to "passthrough" but a site was missed.
- Don't try to post an empty comment if the text is empty but we have inlines; this avoids popping a "you can't post an empty comment" error.
Test Plan: Made an empty comment with an inline in Pholio.
Reviewers: chad
Reviewed By: chad
CC: aran
Maniphest Tasks: T2650
Differential Revision: https://secure.phabricator.com/D5240
Summary:
Currently, there's no easy way for me to tell a user "run this code from the webserver and tell me what it says". Sometimes installs can add new .php files to, e.g., `webroot/rsrc/`, but this is setup-dependent and not universal. Generally I resort to saying "put this into index.php", but that's error prone and not acceptable on active installs.
Add a "debug" controller so I can instead say "put this into support/debug.php, then visit /debug/".
Test Plan: Visited /debug/ with and without support/debug.php files. Visited /staus/.
Reviewers: vrana, btrahan
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D5212
Summary: Deleted code which used channel. Created PhabricatorChatLogChannelQuery.php
Test Plan: By manually checking in the chatlog application.
Reviewers: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D5010
Summary:
- In stack traces, a `,` should clearly be a `.`.
- In Calendar, a 'td' got swapped with a 'p' somewhere.
- In old-style transaction views, strlen() is no longer a sufficient test.
Test Plan:
- Verified stack traces render correctly.
- Verified calendar renders correctly.
- Verified Maniphest transactions with no comment no longer have a little empty div a few pixels high.
Reviewers: chad
Reviewed By: chad
CC: aran
Differential Revision: https://secure.phabricator.com/D4971
Summary: Sgrepped for `"=~/</"` and manually changed every HTML.
Test Plan: This doesn't work yet but it is hopefully one of the last diffs before Phabricator will be undoubtedly HTML safe.
Reviewers: epriestley
CC: aran, Korvin
Maniphest Tasks: T2432
Differential Revision: https://secure.phabricator.com/D4927
Summary:
This is pretty brutal and it adds some `phutil_safe_html()`.
But it is a big step in the right direction.
Test Plan: None.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2432
Differential Revision: https://secure.phabricator.com/D4905
Summary:
This resolves lots of double escaping.
We changed most of `phutil_render_tag(, , $s)` to `phutil_tag(, , $s)` which means that `$s` is now auto-escaped.
Also `pht()` auto escapes if it gets `PhutilSafeHTML`.
Test Plan: None.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2432
Differential Revision: https://secure.phabricator.com/D4889
Summary:
When a developer changes CSS, it is normally sufficient to reload the page to get changes to show up, because browsers revalidate resources on reload.
However, if you reload the page and then an Ajax request adds new CSS to the page, this CSS does not trigger revalidation. The developer must currently clear their cache or re-run `scripts/celerity_mapper.php webroot`, to get this request to skip cache. We rarely use CSS over Ajax right now, so this hasn't cropped up much, but Conpherence does use this and clearing the resource is a big pain.
This seems to work fine normally, but I'm worried it might break some of the extra-celerity-resources stuff Facebook is doing.
Test Plan: In development mode, changed `conpherence/message-pane.css` and saw changes reflected on reload. Verified normal page loads do not cause additional HTTP requests. This change has no effect in production mode.
Reviewers: edward, vrana, btrahan
Reviewed By: vrana
CC: aran
Maniphest Tasks: T2428
Differential Revision: https://secure.phabricator.com/D4902
Summary:
- Separate the ideas of "requested" (explicit user request) vs "started" (user request or sampling).
- Move this code out of index.php into the XHProf stuff (general effort to make index.php smaller).
Test Plan:
Verified that profiling still works, and profiling extends to ajax requests.
Set sampling rate to 2, saw 50% samples.
Looked at database, saw sampling data populating properly.
Reviewers: vrana, nh
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D4891
Summary:
Lots of killed `phutil_escape_html()`.
Done by searching for `AphrontTableView` and then `$rows` (usually) backwards.
Test Plan:
Looked at homepage.
echo id(new AphrontTableView(array(array('<'))))->render();
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4884
Summary:
Done by searching for `AphrontDialogView` and then `appendChild()`.
Also added some `pht()`.
Test Plan: None.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4882
Summary: Searched for `AphrontErrorView` and then for `setTitle()`.
Test Plan: None.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2432
Differential Revision: https://secure.phabricator.com/D4880
Summary: Route all `$_SERVER['HTTP_...']` stuff through AphrontRequest (it would be nice to make this non-static, but the stack is a bit tangled right now...)
Test Plan: Verified CSRF and cascading profiling. `var_dump()`'d User-Agent and Referer and verified they are populated and returned correct values when accessed. Restarted server to trigger setup checks.
Reviewers: vrana
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D4888
Summary: If a page is profiled, add an "X-Phabricator-Profiler" header to all Ajax requests, and profile those too.
Test Plan: Profiled a page, checked Darkconsole, saw profiles for everything.
Reviewers: vrana, btrahan
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D4885
Summary: Done by searching for `AphrontErrorView` and then `appendChild()`.
Test Plan:
Looked at Commit Detail.
Looked at Revision Detail.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Maniphest Tasks: T2432
Differential Revision: https://secure.phabricator.com/D4843
Summary: Refactor options related to verbose error reporting and forcing disk reads into a single developer option.
Test Plan: Run Phabricator with the developer-mode option set and check that errors print stack traces, static assets are always reloaded, etc.
Reviewers: epriestley
Reviewed By: epriestley
CC: aran, Korvin
Differential Revision: https://secure.phabricator.com/D4780
Summary:
This accomplishes three major goals:
# Fixes phutil_render_tag -> phutil_tag callsites in DarkConsole.
# Moves the Ajax request log to a new panel on the left. This panel (and the tabs panel) get scrollbars when they get large, instead of making the page constantly scroll down.
# Loads the panel content over ajax, instead of dumping it into the page body / ajax response body. I've been planning to do this for about 3 years, which is why the plugins are architected the way they are. This should make debugging easier by making response bodies not be 50%+ darkconsole stuff.
Additionally, load the plugins dynamically (the old method predates library maps and PhutilSymbolLoader).
Test Plan:
{F30675}
- Switched between requests and tabs, reloaded page, saw same tab.
- Used "analyze queries", "profile page", triggered errors.
- Verified page does not load anything by default if dark console is closed with Charles.
- Generally banged on it a bit.
Reviewers: vrana, btrahan, chad
Reviewed By: vrana
CC: aran
Maniphest Tasks: T2432
Differential Revision: https://secure.phabricator.com/D4692
Summary:
If you run this code:
json_encode(array('tag' => phutil_tag('div', array())));
...you get this result, because json_encode() does not call toString() on objects:
{"tag":{}}
Instead, convert such objects to their underlying strings. Javelin has support for JX.HTML and for implicit conversion (which is kind of sketchy for other reasons) but it's sort of complicated (only happens on Ajax, not behaviors) and messy (not metadata-based), so ignore it for now.
We'll need to do something similar for serialization to the database. My plan there is just to throw on any objects. The only time we put HTML in the database is cache-related and those tiny number of callsites can manually handle it.
Test Plan: Various ajax things now receive the correct data.
Reviewers: vrana
Reviewed By: vrana
CC: aran
Maniphest Tasks: T2432
Differential Revision: https://secure.phabricator.com/D4684
Summary:
- Implements `javelin_tag()`, which is `javelin_render_tag()` on top of `phutil_tag()` instead of `phutil_render_tag()`.
- Manually converts all or almost all of the trivial callsites.
Test Plan:
- Site does not seem any more broken than before.
Reviewers: vrana
Reviewed By: vrana
CC: aran
Differential Revision: https://secure.phabricator.com/D4639
