In some cases, between Discourse forums the hostname of a URL could match if they are hosting S3 files on the same bucket but the S3 bucket path might not. So e.g. https://testbucket.somesite.com/testpath/some/file/url.png vs https://testbucket.somesite.com/prodpath/some/file/url.png. So has_been_uploaded? was returning true for the second URL, even though it may have been uploaded on a different Discourse forum.
This is a very rare case but must be accounted for, because this impacts UrlHelper.is_local which mistakenly thinks the file has already been downloaded and thus allows the URL to be cooked, where we want to return the full URL to be downloaded using PullHotlinkedImages.
Previously we would raise a warning in the logs if downloading
a file (from s3) takes longer than 60 seconds.
At scale this happens reasonably frequently.
1. Raised the duration to 3 minutes
2. Pulled the resizing mutex out of the downloading mutex
so we have less and clearer error logs
We have the `# frozen_string_literal: true` comment on all our
files. This means all string literals are frozen. There is no need
to call #freeze on any literals.
For files with `# frozen_string_literal: true`
```
puts %w{a b}[0].frozen?
=> true
puts "hi".frozen?
=> true
puts "a #{1} b".frozen?
=> true
puts ("a " + "b").frozen?
=> false
puts (-("a " + "b")).frozen?
=> true
```
For more details see: https://samsaffron.com/archive/2018/02/16/reducing-string-duplication-in-ruby
The `uplaods:migrate_to_s3` rake task should always use the environment variables, because you usually don't want to break your site's uploads during the migration. But restoring a backup should work with site settings as well as environment variables, otherwise you can't restore uploads to S3 from the web interface.
The rake task aborted the migration with "Already migrated" when all upload URLs linked to the correct S3 bucket even though the files didn't exist on S3. By removing the first check we force the rake task to check for the existance of uploads on S3.
A race condition issue is possible when multiple thread/processes are calling this method.
`ls` prints out to stderr "cannot access '...': No such file or directory" if any of the files it's currently trying to list are being removed by the `xargs rm -rf` in an another process. That doesn't affect the result, but it did raise an error before this change.
Tested on a production instance where the original issue was observed.
Co-Authored-By: Régis Hanol <regis@hanol.fr>
### General Changes and Duplication
* We now consider a post `with_secure_media?` if it is in a read-restricted category.
* When uploading we now set an upload's secure status straight away.
* When uploading if `SiteSetting.secure_media` is enabled, we do not check to see if the upload already exists using the `sha1` digest of the upload. The `sha1` column of the upload is filled with a `SecureRandom.hex(20)` value which is the same length as `Upload::SHA1_LENGTH`. The `original_sha1` column is filled with the _real_ sha1 digest of the file.
* Whether an upload `should_be_secure?` is now determined by whether the `access_control_post` is `with_secure_media?` (if there is no access control post then we leave the secure status as is).
* When serializing the upload, we now cook the URL if the upload is secure. This is so it shows up correctly in the composer preview, because we set secure status on upload.
### Viewing Secure Media
* The secure-media-upload URL will take the post that the upload is attached to into account via `Guardian.can_see?` for access permissions
* If there is no `access_control_post` then we just deliver the media. This should be a rare occurrance and shouldn't cause issues as the `access_control_post` is set when `link_post_uploads` is called via `CookedPostProcessor`
### Removed
We no longer do any of these because we do not reuse uploads by sha1 if secure media is enabled.
* We no longer have a way to prevent cross-posting of a secure upload from a private context to a public context.
* We no longer have to set `secure: false` for uploads when uploading for a theme component.
This PR introduces a new secure media setting. When enabled, it prevent unathorized access to media uploads (files of type image, video and audio). When the `login_required` setting is enabled, then all media uploads will be protected from unauthorized (anonymous) access. When `login_required`is disabled, only media in private messages will be protected from unauthorized access.
A few notes:
- the `prevent_anons_from_downloading_files` setting no longer applies to audio and video uploads
- the `secure_media` setting can only be enabled if S3 uploads are already enabled and configured
- upload records have a new column, `secure`, which is a boolean `true/false` of the upload's secure status
- when creating a public post with an upload that has already been uploaded and is marked as secure, the post creator will raise an error
- when enabling or disabling the setting on a site with existing uploads, the rake task `uploads:ensure_correct_acl` should be used to update all uploads' secure status and their ACL on S3
POSIX's `head` specification states: "The application shall ensure that the number option-argument is a positive decimal integer"
Negative values are supported on GNU `head`, so this works in the discourse docker image. However, in some environments (e.g. macOS), the system `head` version fails with a negative `n` parameter.
This commit does two things:
Checks the status at each stage of the pipe, so it cannot fail silently
Flip the `ls` command to list in descending time order, and use `tail -n +501` instead of `head -n -500`.
The visible result is that macOS users no longer see head: illegal line count -- -500 printed throughout the test suite.
Doing .pluck(:column).first is a very common pattern in Discourse and in
most cases, a limit cause isn't being added. Instead of adding a limit
clause to all these callsites, this commit adds two new methods to
ActiveRecord::Relation:
pluck_first, equivalent to limit(1).pluck(*columns).first
and pluck_first! which, like other finder methods, raises an exception
when no record is found
* Support private uploads in S3
* Use localStore for local avatars
* Add job to update private upload ACL on S3
* Test multisite paths
* update ACL for private uploads in migrate_to_s3 task
Net::HTTP always returns ASCII-8BIT encoding. File.read auto-detects the encoding. This leads to an encoding inconsistency between a fresh download, and a cached download. This commit ensures all downloaded files are treated equally, by always returning the cached version from the filesystem, even during initial download.
One symptom of this problem is during theme exports: https://meta.discourse.org/t/116907
Related ruby ticket: https://bugs.ruby-lang.org/issues/2567
This reduces chances of errors where consumers of strings mutate inputs
and reduces memory usage of the app.
Test suite passes now, but there may be some stuff left, so we will run
a few sites on a branch prior to merging
When the S3 store was enabled, we were only applying the S3 CDN.
So all images stored locally, like the emojis, were never put on the local CDN.
Fixed a bunch of CookedPostProcessor test by adding a call to 'optimize_urls'
in order to get final URLs.
I also removed the unnecessary PrettyText.add_s3_cdn method since this is already
handled in the CookedPostProcessor.
This refactors handling of s3 so it can be specified via GlobalSetting
This means that in a multisite environment you can configure s3 uploads
without actual sites knowing credentials in s3
It is a critical setting for situations where assets are mirrored to s3.
FileUtils.move actually ends up raising an "unknown file type" error if the file doesn't exist instead of Errno::ENOENT.
It's possible to rescue this, but in the end it's easier to just ask move not to throw an error, since we're going to throw it away anyway.
