Summary: Fixes T8277. Transforming files can race; resolve the race after we lose.
Test Plan:
- Added `sleep(10)` near the bottom of the transform controller.
- Transformed a file in two browser windows at the same time; got something like this (exception corresponds to the loser of the race):
{F412526}
- Applied patch.
- Repeated process, got this:
{F412527}
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T8277
Differential Revision: https://secure.phabricator.com/D12965
Summary: Fixes T5703. These have been unused in production for a while and the new stuff seems good.
Test Plan: Mostly `grep`.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T5703
Differential Revision: https://secure.phabricator.com/D12949
Summary: Remove the `PhabricatorDefaultSearchEngineSelector` class. This is quite similar to D12053.
Test Plan: Went to `/view/PhabricatorSearchApplication/` and saw the storage engine configuration. Set `search.elastic.host` and saw the highlighted storage engine change.
Reviewers: #blessed_reviewers, epriestley
Reviewed By: #blessed_reviewers, epriestley
Subscribers: Korvin, epriestley
Differential Revision: https://secure.phabricator.com/D12670
Summary: Ref T6403. This was actually simple stuff.
Test Plan: changed the edit policy of a paste. changed the edit and join policy of a phame blog.
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin, epriestley
Maniphest Tasks: T6403
Differential Revision: https://secure.phabricator.com/D12933
Summary:
Fixes T6956. Before this change, we called PhabricatorUser::getOmnipotentUser in the various delete methods to query the data. Now, we use $engine->getViewer(), since its always a good thing to have less calls to PhabricatorUser::getOmnipotentUser thrown around the codebase.
I used the "codemod" tool to audit the existing calls to PhabricatorDestructorEngine (all of them) so ostensibly this gets all the spots. If I missed something though, its still going to work, so this change is very low risk.
Test Plan: ./bin/remove destroy P1; visit P1 and get a 404
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin, epriestley
Maniphest Tasks: T6956
Differential Revision: https://secure.phabricator.com/D12866
Summary: Use `__CLASS__` instead of hard-coding class names. Depends on D12605.
Test Plan: Eyeball it.
Reviewers: #blessed_reviewers, epriestley
Reviewed By: #blessed_reviewers, epriestley
Subscribers: hach-que, Korvin, epriestley
Differential Revision: https://secure.phabricator.com/D12806
Summary:
Ref T7707. The general form of this can probably be refined somewhat over time as we have more use cases.
I put this cache on the user object itself because we essentially always need this data and it's trivial to invalidate the cache (we can do it implicilty during reads).
Also fix an issue with short, wide images not thumbnailing properly after recent changes.
Test Plan:
- Loaded some pages; saw caches write; saw good pictures.
- Reloaded; saw cache reads; saw good pictures.
- Changed profile picture; saw immediate update.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7707
Differential Revision: https://secure.phabricator.com/D12826
Summary:
Ref T7707. Fixes T7879. Fixes T4406. When creating profile images:
- Use the new transforms;
- mark them as "profile" images so they're forced to the most-open policies.
Test Plan:
- Set restrictive default file policies.
- Changed profile picture, project pictures, etc. Verified they were visible to logged-out users.
- Registered via OAuth.
- Updated a Conpherence thread image.
- Browsed around looking for profile images, fixed sizing on everything I could find.
Reviewers: chad, btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7879, T7707, T4406
Differential Revision: https://secure.phabricator.com/D12821
Summary:
Ref T7707. This ends up being sort of complicated: to support 100x100 images in T4406, we need to scale small images //up// so they look OK when we scale them back down with `background-size` in CSS.
The rest of it is mostly straightforward.
Test Plan:
- Did an OAuth handshake and saw a scaled-up, scaled-down profile picture that looked correct.
- Used Pholio, edited pholio, embedded pholio.
- Uploaded a bunch of small/weird/big images and regenerated all their transforms.
- Uploaded some text files into Pholio.
- Grepped for removed methods, etc.
Reviewers: chad, btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7707
Differential Revision: https://secure.phabricator.com/D12818
Summary: Ref T7707. Move the 220px (file uploads) and 100px (Pholio thumbgrid) previews over to the new stuff.
Test Plan: Uploaded a bunch of images to remarkup and Pholio; they generated reasonable results in the web UI.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7707
Differential Revision: https://secure.phabricator.com/D12814
Summary: Ref T7707. For animated GIFs, use imagemagick if it is available.
Test Plan: Generated small versions of a bunch of different GIFs.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7707
Differential Revision: https://secure.phabricator.com/D12813
Summary: Ref T7707. These transforms have a single maximum dimension instead of fixed X and Y dimensions.
Test Plan: Transformed a bunch of files with different sizes/aspect ratios, got sensible results.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7707
Differential Revision: https://secure.phabricator.com/D12812
Summary:
Ref T7707. Ref T4406. Ref T2479. This implements the profile-style (fixed width and height) transforms in a modern way.
- Added a "regnerate" feature to the support UI to make testing easier and surface errors.
- Laboriously check errors from everything.
- Fix the profile thumbnailing so it crops properly instead of leaving margins.
- Also defuses the "gigantic white PNG" attack.
This doesn't handle the imagemagick case (for animated GIFs) yet.
Test Plan:
- Uploaded a variety of wide/narrow/small/large files and converted them into sensible profile pictures.
- Tried to thumbnail some text files.
- Set the pixel-size and file-size limits artificially small and hit them.
- Used "regenerate" a bunch while testing the rest of this stuff.
- Verified that non-regenerate flows still produce a default/placeholder image.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T4406, T2479, T7707
Differential Revision: https://secure.phabricator.com/D12811
Summary: Ref T7707. Fixes T4724. I misread the report on T4724; this is trivial. We're just reading the wrong properties in setting "width" and "height" attributes, the actual thumbnailing logic is fine.
Test Plan: Uploaded image from T4724, saw it have a proper aspect ratio.
Reviewers: btrahan, chad
Reviewed By: chad
Subscribers: spicyj, epriestley
Maniphest Tasks: T4724, T7707
Differential Revision: https://secure.phabricator.com/D12810
Summary:
Ref T7707. Ref T2479. Ref T5258.
The thumbnailing code is some of the only code in the codebase which doesn't use exceptions to handle errors. I'm going to convert it to use exceptions; make sure they do something reasonable at top level.
Strategy here is:
- By default, we just fall back to a placeholder image if anything goes wrong.
- Later, I'll likely add a "debug" workflow from the new "Transforms" UI which will surface the specific exception instead (the code can't really raise any interesting exceptions right now).
Test Plan: Faked an exception and saw some reasonable default images.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T5258, T2479, T7707
Differential Revision: https://secure.phabricator.com/D12809
Summary:
Ref T7707. Available transforms are currently relatively hard-coded and don't really have any support UI.
Modularize them so we can build some support UI.
This doesn't actually //use// any of the new stuff yet: I want to make a clean cutover once I fix the aspect ratio stuff so I can pick up a cachekey/URI change as a side effect.
Test Plan: {F400524}
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: chad, epriestley
Maniphest Tasks: T7707
Differential Revision: https://secure.phabricator.com/D12808
Summary: Fixes T7480, File names should be editable and the event should show up in feed.
Test Plan: Upload a file, view file details, edit file, change file name by adding a space and a word to the name, save changes, file name should retain space and not normalize the name, file details should show the edit event, install feed should correctly show an event for the action.
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: Korvin, epriestley
Maniphest Tasks: T7480
Differential Revision: https://secure.phabricator.com/D12561
Summary: Fixes T7685. This required making the global drag and drop behavior able to "uninstall" itself so to speak, and then it re-installs it self as necessary.
Test Plan:
Did the following all successfully
- uploaded a file to homepage
- homepage -> differential -- no way to upload via drag and drop
- homepage -> differential -> homepage -- uploaded a file
- homepage -> differential -> browser back button to homepage -- uploaded a file
Reviewers: epriestley
Reviewed By: epriestley
Subscribers: Korvin, epriestley
Maniphest Tasks: T7685
Differential Revision: https://secure.phabricator.com/D12534
Summary:
Ref T4100. Ref T5595. These functions are trivial for now, but move us toward being able to define more default query behavior by default.
Future changes will give these methods meaningful, nontrivial behaviors.
Test Plan: `arc unit --everything`
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T5595, T4100
Differential Revision: https://secure.phabricator.com/D12454
Summary:
Ref T4100. Ref T5595.
To support a unified "Projects:" query across all applications, a future diff is going to add a set of "Edge Logic" capabilities to `PolicyAwareQuery` which write the required SELECT, JOIN, WHERE, HAVING and GROUP clauses for you.
With the addition of "Edge Logic", we'll have three systems which may need to build components of query claues: ordering/paging, customfields/applicationsearch, and edge logic.
For most clauses, queries don't currently call into the parent explicitly to get default components. I want to move more query construction logic up the class tree so it can be shared.
For most methods, this isn't a problem, but many subclasses define a `buildWhereClause()`. Make all such definitions protected and consistent.
This causes no behavioral changes.
Test Plan: Ran `arc unit --everything`, which does a pretty through job of verifying this statically.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: yelirekim, hach-que, epriestley
Maniphest Tasks: T4100, T5595
Differential Revision: https://secure.phabricator.com/D12453
Summary:
Ref T7803. Ref T5873. I want to drive Conduit through more shared infrastructure, but can't currently add parameters automatically.
Put a `getX()` around the `defineX()` methods so the parent can provide default behaviors.
Also like 60% of methods don't define any special error types; don't require them to implement this method. I want to move away from this in general.
Test Plan:
- Ran `arc unit --everything`.
- Called `conduit.query`.
- Browsed Conduit UI.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: hach-que, epriestley
Maniphest Tasks: T5873, T7803
Differential Revision: https://secure.phabricator.com/D12380
Summary:
Ref T7803. Some Query subclasses implement getPagingColumn() in a trivial way, usually to provide a table alias.
Formalize the concept of a primary table alias, and remove obsoleted getPagingColumn() implementations.
Test Plan: Issued affected queries.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7803
Differential Revision: https://secure.phabricator.com/D12356
Summary: These arrays looks a little odd, most likely due to the autofix applied by `ArcanistXHPASTLinter::LINT_ARRAY_SEPARATOR`. See D12296 in which I attempt to improve the autocorrection from this linter rule.
Test Plan: N/A
Reviewers: epriestley, #blessed_reviewers
Reviewed By: epriestley, #blessed_reviewers
Subscribers: epriestley
Differential Revision: https://secure.phabricator.com/D12281
Summary:
Fixes T7199. This still isn't a shining example of perfect code, but the raw amount of copy/paste is much lower than it used to be.
- Reduce code duplication between existing receivers.
- Expose receiving objects in help menus where appropriate.
- Connect some "TODO" receivers.
Test Plan:
- Sent mail to every supported object type.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7199
Differential Revision: https://secure.phabricator.com/D12249
Summary: Ref T7199. Everyone can have a mail command! You can have a mail command! You can have a mail command! Mail commands for everyone!
Test Plan: Used `bin/mail receive-test` to issue commands against files and pastes.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7199
Differential Revision: https://secure.phabricator.com/D12238
Summary: Ref T7199. Half of these aren't even reachable, but make some progress toward reducing the amount of nonsense and garbage in mail handling.
Test Plan: Tested all reachable handlers with `bin/mail receive-test`.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7199
Differential Revision: https://secure.phabricator.com/D12237
Summary:
Ref T7199. This prepares for an exciting new world of more powerful "!action" commands. In particular:
- We parse multiple commands per mail.
- We parse command arguments (these are currently not used).
- We parse commands at the beginning or end of mail.
Additionally:
- Do a quick modernization pass on all handlers.
- Break legacy compatibility with really hacky Facebook stuff (see T1992). They've theoretically been on notice for a year and a half, and their setup relies on calling very old reply handler APIs directly.
- Some of these handlers had some copy/paste fluff.
- The Releeph handler is unreachable, but fix it //in theory//.
Test Plan:
- Sent mail to a file; used "!unsubscribe".
- Sent mail to a legalpad document; used "!unsubscribe".
- Sent mail to a task; used various "!close", "!claim", "!assign", etc.
- Sent mail to a paste.
- Sent mail to a revision; used various "!reject", "!claim", etc.
- Tried to send mail to a pull request but it's not actually reachable.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7199
Differential Revision: https://secure.phabricator.com/D12230
Summary:
Ref T7199. Although this is useful for discovery, it's un-useful enough that we already have an option to disable it, and most applications do not provide any meaningful instructions.
Throwing it away makes it easier to move forward and lets us get rid of a config option.
This is becoming a more advanced/power-user feature anyway, and the new syntax will be significantly more complex and hard to explain with a one-liner. I'm currently thinking that I'll maybe make the "help" menu a dropdown and give it some options like:
+---+
| O |
+---+---------------------+
| Maniphest Documentation |
| Maniphest Email Actions |
+-------------------------+
Then you click the "Email Actions" thing and get a runtime-derived list of available options. Not sure if I'll actually build that, but I think we can fairly throw the in-mail instructions away even if we don't go in that specific direction.
Test Plan: Grepped for `replyHandlerInstructions`, got no hits.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7199
Differential Revision: https://secure.phabricator.com/D12229
Summary: Ref T7199. Guess no one has ever tried to reply to file mail.
Test Plan: Used `bin/mail receive-test` to send mail to files.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7199
Differential Revision: https://secure.phabricator.com/D12228
Summary: Ref T7689. Ref T4100. This advances the goals of removing `loadViewerHandles()` (only 67 callsites remain!) and letting tokenizers some day take token functions like `viewer()` and `members(differential)`.
Test Plan:
- Sent a new message; used "To".
- I simplified the cancel URI construction slightly because it's moot in all normal cases.
- Edited a thread; used "Add Participants".
- Searched rooms; used "Participants".
- Searched countdowns; used "Authors".
- Created a diff; used "Repository".
- Edited a revision; edited "Projects"; edited "Reveiwers"; edited "Subscribers".
- Searched for revisions; edited "responsible users"; "authors"; "reviwers"; "subscribers"; "repositories".
- Added revision comments; edited "Add Reveiwers"; "Add Subscribers".
- Commented on a commit; edited "Add Auditors"; "Add subscribers".
- Edited a commit; edited "Projects".
- Edited a repository; edited "Projects".
- Searched feed, used "include Users"; "include Proejcts".
- Searched files, used "authors".
- Edited initiative; edited "Projects".
- Searched backers; used "Backers".
- Searched initiatives; used "Owners".
- Edited build plans; edited "Run Command".
- Searched Herald; used "Authors".
- Added signature exemption in Legalpad.
- Searhced legalpad; used "creators"; used "contributors".
- Searched signatures; used "documents"; used "signers".
- Created meme.
- Searched macros; used "Authors".
- Used "Projects" in Maniphest reports.
- Used Maniphest comment actions.
- Edited Maniphest tasks; edited "Assigned To"; edited "CC"; edited "projects".
- Used "parent" in Maniphest task creation workflow.
- Searched for projects; used "assigned to"; "in any projec"; "in all projects"; "not in projects"; "in users' projects"; "authors"; "subscribers".
- Edited Maniphest bug filing domains, used "Default Author".
- Searched for OAuth applications, used "Creators".
- Edited Owners pacakge; edited "Primary Owner"; edited "Owners".
- Searched for Owners packages; used "Owner".
- OMG this UI is OLD
- Edited a paste; edited "Projects".
- Searched for paste; used "Authors".
- Searched user activity log; used "Actors"; used "Users".
- Edited a mock; edited "Projects"; edited "CC".
- Searched for mocks; used "Authors".
- Edited Phortune account; edited "Members".
- Edited Phortune merchant account; edited "Members".
- Searched Phrequent; used "Users".
- Edited Ponder question; sued "projects".
- Searched Ponder; used "Authors"; used "Answered By".
- Added project members.
- Searched for projects; used "Members".
- Edited a Releeph product; edited "Pushers".
- Searched pull requests; searched "Requestors".
- Edited an arcanist project; used "Uses Symbols From".
- Searhced push logs; used "Repositories"; used "Pushers".
- Searched repositories; used "In nay project".
- Used global search; used Authors/owners/Subscribers/In Any Project.
- Edited a slowvote; used "Projects".
- Searched slovotes; used "Authors".
- Created a custom "Users" field; edited and searched for it.
- Made a whole lot of typos in this list. ^^^^^^
Did not test:
- Lint is nontrivial to test locally, I'll test it in production.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T4100, T7689
Differential Revision: https://secure.phabricator.com/D12224
Summary: Ref T6755. I'll add some notes there about specifics.
Test Plan:
- Made connections to HTTP and HTTPS URIs.
- Added some debugging code to verify that HTTP URIs were pre-resolved.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T6755
Differential Revision: https://secure.phabricator.com/D12169
Summary:
Ref T6755. Although we do not return response bodies, it is possible to perform crude portscanning if you can execute a DNS rebinding attack (which, for now, remains theoretical).
Limit users to 60 requests / hour to make it less feasible. This would require ~30 years to portscan all ports on a `/32` netblock.
Users who can guess that services may exist can confirm their existence more quickly than this, but if the attacker already had a very small set of candidate services it seems unlikely that portscanning would be of much use in executing the attack.
This protection should eventually be applied to T4190, too (that task also has other considerations).
Test Plan: Set rate limit very low, hit rate limit.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T6755
Differential Revision: https://secure.phabricator.com/D12168
Summary:
Ref T6755. This mitigates an attack where you:
- compromise an administrative account;
- configure "text/plain" as an "image" MIME type; and
- create a new macro sourced from a sensitive resource which is locally accessible over HTTP GET, using DNS rebinding.
You can then view the content of the resource in Files. By preventing the compromised account from reconfiguring the MIME types, the server will instead destroy the response and prevent the attacker from seeing it.
In general, these options should change very rarely, and they often sit just beyond the edge of security vulnerabilities anyway.
For example, if you ignore the warnings about an alternate file domain and elect to serve content from the primary domain, it's still somewhat difficult for an attacker to exploit the vulnerability. If they can add "text/html" or "image/svg+xml" as image MIME types, it becomes trivial. In this case not having an alternate domain is the main issue, but easy modification of this config increases risk/exposure.
Test Plan: Viewed affected config and saw that it is locked.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T6755
Differential Revision: https://secure.phabricator.com/D12154
Summary:
Ref T6755. This improves our resistance to SSRF attacks:
- Follow redirects manually and verify each component of the redirect chain.
- Handle authentication provider profile picture fetches more strictly.
Test Plan:
- Tried to download macros from various URIs which issued redirects, etc.
- Downloaded an actual macro.
- Went through external account workflow.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T6755
Differential Revision: https://secure.phabricator.com/D12151
Summary: This signature changed at some point after I tested things and I didn't catch it.
Test Plan: Destroyed a chunked large file with `bin/remove`.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Differential Revision: https://secure.phabricator.com/D12152
Summary:
Ref T6755. This is a partial fix, but:
- Allow netblocks to be blacklisted instead of making the feature all-or-nothing.
- Default to disallow requests to all reserved private/local/special IP blocks. This should generally be a "safe" setting.
- Explain the risks better.
- Improve the errors rasied by Macro when failing.
- Removed `security.allow-outbound-http`, as it is superseded by this setting and is somewhat misleading.
- We still make outbound HTTP requests to OAuth.
- We still make outbound HTTP requests for repositories.
From a technical perspective:
- Separate URIs that are safe to link to or redirect to (basically, not "javascript://") from URIs that are safe to fetch (nothing in a private block).
- Add the default blacklist.
- Be more careful with response data in Macro fetching, and don't let the user see it if it isn't ultimately valid.
Additionally:
- I want to do this check before pulling repositories, but that's enough of a mess that it should go in a separate diff.
- The future implementation of T4190 needs to perform the fetch check.
Test Plan:
- Fetched a valid macro.
- Fetched a non-image, verified it didn't result in a viewable file.
- Fetched a private-ip-space image, got an error.
- Fetched a 404, got a useful-enough error without additional revealing response content (which is usually HTML anyway and not useful).
- Fetched a bad protocol, got an error.
- Linked to a local resource, a phriction page, a valid remote site, all worked.
- Linked to private IP space, which worked fine (we want to let you link and redierect to other private services, just not fetch them).
- Added and executed unit tests.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T6755
Differential Revision: https://secure.phabricator.com/D12136
Summary:
Fixes T7621. The engine selection code started out making sense, but didn't make as much sense by the time I was done with it.
Specifically, from the vanilla file upload, we may incorrectly try to write directly to the chunk storage engine. This is incorrect, and produces a confusing/bad error.
Make chunk storage engines explicit and don't try to do single-file one-shot writes to them.
Test Plan:
- Tried to upload a large file with vanilla uploader, got better error message.
- Uploaded small and large files with drag and drop.
- Viewed {nav Files > Help/Options}.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7621
Differential Revision: https://secure.phabricator.com/D12110
Summary:
Ref T7149. Currently, global drag and drop always uses the most open visibility policy on the install. This was appropriate before the application preference was introduced, but default to the application preference now.
In particular, this supports a default value of "Administrators" in the Phacility cluster.
Also simplify/clean up some code.
Test Plan:
- Set application default policy to "Adminstrators".
- Uploaded file via drag-and-drop, saw "administrators" policy.
- Uploaded file via `arc upload`, saw "administrators" policy.
- Saw better URI for a text file upload after patch.
- Uploaded file via drag-and-drop-to-textarea, saw "only viewer" policy.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7149
Differential Revision: https://secure.phabricator.com/D12093
Summary:
Ref T7149. We can simplify configuration somewhat by removing the upload limit setting, now that we support arbitrarily large files.
- Merge configuration documentation.
- Tell users to set things to at least 32MB. This is 8MB maximum one-shot file + 4x headroom. Chunk sizes are 4MB.
Test Plan:
- Faked all the setup warnings.
- Read documentation.
- Uploaded some files.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7149
Differential Revision: https://secure.phabricator.com/D12083
Summary: Ref T7149. This works now, so enable it.
Test Plan:
- Uploaded large and small files in Firefox, Safari and Chrome.
- Uploaded large files with `arc upload`.
- Stopped/resumed large files with all clients.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7149
Differential Revision: https://secure.phabricator.com/D12079
Summary: Ref T7149. This was just to make testing easier, but chunking substantially works now.
Test Plan: `grep`
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: epriestley
Maniphest Tasks: T7149
Differential Revision: https://secure.phabricator.com/D12076
Summary: Ref T7149. Return a real iterator from the Chunk engine, which processes chunks sequentially.
Test Plan:
This is a bit hard to read, but shows the underlying chunks being accessed one at a time and only some being accessed when requesting a range of a file:
```
$ ./bin/files cat F878 --trace --begin 100 --end 256
...
>>> [10] <query> SELECT * FROM `file_storageblob` WHERE `id` = 85
<<< [10] <query> 240 us
better software.
Phabricat>>> [11] <query> SELECT * FROM `file_storageblob` WHERE `id` = 84
<<< [11] <query> 205 us
or includes applications for:
>>> [12] <query> SELECT * FROM `file_storageblob` WHERE `id` = 83
<<< [12] <query> 226 us
- reviewing and auditing source>>> [13] <query> SELECT * FROM `file_storageblob` WHERE `id` = 82
<<< [13] <query> 203 us
code;
- hosting and browsing >>> [14] <query> SELECT * FROM `file_storageblob` WHERE `id` = 81
<<< [14] <query> 231 us
repositories;
- tracking bugs;
```
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: joshuaspence, epriestley
Maniphest Tasks: T7149
Differential Revision: https://secure.phabricator.com/D12073
Summary:
Ref T7149. This still buffers the whole file, but is reaaaaal close to not doing that.
Allow Responses to be streamed, and rewrite the range stuff in the FileResponse so it does not rely on having the entire content available.
Test Plan:
- Artificially slowed down downloads, suspended/resumed them (works in chrome, not so much in Safari/Firefox?)
- Played sounds in Safari/Chrome.
- Viewed a bunch of pages and files in every browser.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: joshuaspence, epriestley
Maniphest Tasks: T7149
Differential Revision: https://secure.phabricator.com/D12072
Summary: Ref T7149. A couple diffs down the line, this will let us emit chunked files without doing all the work up front or holding the entire file in RAM.
Test Plan:
(Some newlines added for clarity.)
```
$ ./bin/files cat F942
ABCDEFGHIJKLMNOPQRSTUVWXYZ
$ ./bin/files cat F942 --begin 1
BCDEFGHIJKLMNOPQRSTUVWXYZ
$ ./bin/files cat F942 --end 10
ABCDEFGHIJ
$ ./bin/files cat F942 --begin 3 --end 5
DE
$
```
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: joshuaspence, epriestley
Maniphest Tasks: T7149
Differential Revision: https://secure.phabricator.com/D12071
Summary: Ref T7149. We can't compute hashes of large files efficiently, but we can resume uploads by the same author, with the same name and file size, which are only partially completed. This seems like a reasonable heuristic that is unlikely to ever misfire, even if it's a little magical.
Test Plan:
- Forced chunking on.
- Started uploading a chunked file.
- Closed the browser window.
- Dropped it into a new window.
- Upload resumed //(!!!)//
- Did this again.
- Downloaded the final file, which successfully reconstructed the original file.
Reviewers: btrahan
Reviewed By: btrahan
Subscribers: joshuaspence, chad, epriestley
Maniphest Tasks: T7149
Differential Revision: https://secure.phabricator.com/D12070