Alexander Ebert [Fri, 15 Oct 2021 15:23:06 +0000 (17:23 +0200)]
Incorrect data type used in AJAX requests for search requests
The `Set` type is not understood by the browser's AJAX API and thus silently discarded.
Tim Düsterhus [Thu, 14 Oct 2021 13:17:37 +0000 (15:17 +0200)]
Merge branch '5.3' into 5.4
Tim Düsterhus [Thu, 14 Oct 2021 13:14:54 +0000 (15:14 +0200)]
Merge branch '5.2' into 5.3
Tim Düsterhus [Thu, 14 Oct 2021 13:11:44 +0000 (15:11 +0200)]
Merge branch '3.1' into 5.2
Tim Düsterhus [Thu, 14 Oct 2021 13:10:10 +0000 (15:10 +0200)]
Fix EmailNewActivationCodeForm
This got broken, because it inherits from RegisterNewActivationForm and the “is
already enabled” validation was moved into a dedicated method within there. This
is a perfect example of why one should never inherit from controllers …
see
f394421c0cc7e8879007092e40e540b2fd1118c1
Tim Düsterhus [Thu, 14 Oct 2021 13:03:35 +0000 (15:03 +0200)]
Fix bad merge from 5.3 to 5.4
Commit
f394421c0cc7e8879007092e40e540b2fd1118c1 was incorrectly applied.
Tim Düsterhus [Wed, 13 Oct 2021 14:21:09 +0000 (16:21 +0200)]
Allow unblocking non-blockable users from within the profile
Fixes #4548
WoltLab [Wed, 13 Oct 2021 10:37:41 +0000 (10:37 +0000)]
Updating minified JavaScript files
Tim Düsterhus [Wed, 13 Oct 2021 10:36:21 +0000 (12:36 +0200)]
Update npm dependencies in extra/
WoltLab [Wed, 13 Oct 2021 10:31:37 +0000 (10:31 +0000)]
Updating minified JavaScript files
Marcel Werk [Fri, 8 Oct 2021 14:01:18 +0000 (16:01 +0200)]
Support for the embedding of private vimeo videos
Tim Düsterhus [Tue, 5 Oct 2021 07:00:18 +0000 (09:00 +0200)]
Merge pull request #4539 from WoltLab/php-ddl-reject-duplicate-index-column
Reject indices with duplicate columns in DatabaseTableChangeProcessor
Tim Düsterhus [Mon, 4 Oct 2021 14:32:53 +0000 (16:32 +0200)]
Merge pull request #4540 from WoltLab/5.4-image-proxy-exception
Correctly wrap \RuntimeException from body reading in \DomainExceptio…
joshuaruesweg [Mon, 4 Oct 2021 14:28:44 +0000 (16:28 +0200)]
Correctly wrap \RuntimeException from body reading in \DomainException in ImageProxyAction
Tim Düsterhus [Mon, 4 Oct 2021 14:04:11 +0000 (16:04 +0200)]
Reject indices with duplicate columns in DatabaseTableChangeProcessor
Resolves #4536
Joshua Rüsweg [Mon, 4 Oct 2021 13:31:19 +0000 (15:31 +0200)]
Merge pull request #4538 from WoltLab/5.4-disable-unfurling-in-signatures
Disable unfurled urls in signatures
joshuaruesweg [Mon, 4 Oct 2021 12:53:14 +0000 (14:53 +0200)]
Disable unfurled urls in signatures
Tim Düsterhus [Wed, 29 Sep 2021 13:38:50 +0000 (15:38 +0200)]
Fix check whether a non-owned FOREIGN KEY is being dropped in DatabaseTableChangeProcessor
The reproducer and fix is effectively identical to the one in
167291206e57ffb9bc043308682061e5e499ff45.
Package A: Installs FOREIGN KEY (someOtherUserID) REFERENCES wcf1_user (userID)
Package B: Installs FOREIGN KEY (userID) REFERENCES wcf1_user (userID)
Package B: Drops FOREIGN KEY (userID) REFERENCES wcf1_user (userID)
It was erroneously detected that Package B would drop the foreign key owned by
Package A, but possibly only after the foreign key has already been (correctly)
dropped. This delay in check is caused by the `continue 2;` skipping any other
foreign keys after matching up one foreign key.
The actual dropping logic was already correct, just the safety check was
incorrect.
see #4434
Alexander Ebert [Wed, 29 Sep 2021 12:06:46 +0000 (14:06 +0200)]
Force blur the editor after replying with a message
See https://community.woltlab.com/thread/292195-probleme-mit-opera-mobile-unter-android/
Alexander Ebert [Wed, 29 Sep 2021 12:06:09 +0000 (14:06 +0200)]
Force blur the editor after replying with a message
See https://community.woltlab.com/thread/292195-probleme-mit-opera-mobile-unter-android/
Joshua Rüsweg [Wed, 29 Sep 2021 08:46:35 +0000 (10:46 +0200)]
Merge pull request #4532 from WoltLab/unfurl-body-read-failure
Correctly wrap \RuntimeException from body reading in DownloadFailed in UnfurlResponse
Tim Düsterhus [Tue, 28 Sep 2021 15:10:05 +0000 (17:10 +0200)]
Correctly wrap \RuntimeException from body reading in DownloadFailed in UnfurlResponse
Tim Düsterhus [Tue, 28 Sep 2021 14:01:19 +0000 (16:01 +0200)]
Merge branch '5.3' into 5.4
Tim Düsterhus [Tue, 28 Sep 2021 13:58:46 +0000 (15:58 +0200)]
Merge pull request #4531 from WoltLab/http-request-timeout
Configure emergency timeout in HTTPRequest
Tim Düsterhus [Tue, 28 Sep 2021 13:21:30 +0000 (15:21 +0200)]
Merge remote-tracking branch 'origin/5.4' into 5.4
Tim Düsterhus [Tue, 28 Sep 2021 13:18:52 +0000 (15:18 +0200)]
Add explicit check whether the port is numeric in Redis wrapper
This improves error messages.
Tim Düsterhus [Tue, 28 Sep 2021 13:17:30 +0000 (15:17 +0200)]
Merge branch '5.3' into 5.4
Tim Düsterhus [Tue, 28 Sep 2021 13:13:42 +0000 (15:13 +0200)]
Cast the Redis port to int
The `Redis::connect()` method expects the `$port` parameter to be an integer.
PHP will automatically cast numeric strings to an integer, but error out with
an TypeError if the string is not a well-formed number. This TypeError will not
be caught in an `catch(\Exception $e)` block, because TypeError does not
inherit Exception.
Perform an explicit cast to ensure the fallback to DiskCacheSource works.
Tim Düsterhus [Tue, 28 Sep 2021 12:31:33 +0000 (14:31 +0200)]
Configure emergency timeout in HTTPRequest
The connect and read timeouts might not reliably trigger in all cases.
Configure a large overall timeout to ensure PHP workers will terminate
eventually.
see
2dbd5654cb9faff45bb51df9a2f3834bd320cc00
Alexander Ebert [Mon, 27 Sep 2021 15:00:48 +0000 (17:00 +0200)]
Incorrect detection of HTML tags
The previous regex was incorrect and caused false-positive matches. One such case was a `<td>The …</td>` which translated into `###td ###The …`, causing it to be recognized as a `<th>`.
The new regex is much more restrictive by requiring at least one whitespace after the tag name if there is additional content.
joshuaruesweg [Mon, 27 Sep 2021 11:16:31 +0000 (13:16 +0200)]
Fix removing reactions on guests content
Since MySQL 8 the deletion of reactions on contents created by guests might fail. The ReactionHandler tries to update the likesReceived column for a non-existent user, sending the empty string as the userID. Recent versions of MySQL 8 error out with MySQL error 1292. The following MySQL bug appears to be related:
https://bugs.mysql.com/bug.php?id=101806
Tim Düsterhus [Mon, 27 Sep 2021 08:32:14 +0000 (10:32 +0200)]
Merge pull request #4526 from WoltLab/session-cookie-lifetime
Decrease the session cookie lifetime leeway to 1 week
Tim Düsterhus [Mon, 27 Sep 2021 08:03:13 +0000 (10:03 +0200)]
Decrease the session cookie lifetime leeway to 1 week
With the increase of the user session lifetime to 2 months, simply multiplying
by two results in an excessive cookie lifetime.
Decrease this to a constant leeway of 1 week. If the cookie in the browser
expires, the session on the server should be long gone, even for wildly
incorrect local clocks.
Joshua Rüsweg [Mon, 27 Sep 2021 07:50:54 +0000 (09:50 +0200)]
Merge pull request #4525 from WoltLab/session-device-icon
Move Session::getDeviceIcon() into UserAgent::getDeviceIcon()
Tim Düsterhus [Fri, 24 Sep 2021 14:29:03 +0000 (16:29 +0200)]
Move Session::getDeviceIcon() into UserAgent::getDeviceIcon()
This method does not really belong into the Session class.
Tim Düsterhus [Fri, 24 Sep 2021 13:27:48 +0000 (15:27 +0200)]
Validate the XSRF-Token in DeleteSessionAction
This is not necessarily required, because the `sessionID` already contains high
entropy. However the JavaScript code already provides the XSRF-Token, so let's
validate it for completeness.
Tim Düsterhus [Fri, 24 Sep 2021 12:34:39 +0000 (14:34 +0200)]
Remove SECURITY_TOKEN* constants from constants.php
These were effectively deprecated in
3f6a261b1e6a3804370eb1e2a046ea6c666dbedd.
Tim Düsterhus [Fri, 24 Sep 2021 12:33:48 +0000 (14:33 +0200)]
Remove SID* constants from constants.php
These were removed in
8a35fd6de81f1138456fb777eb57d4b3907c0c66.
Alexander Ebert [Fri, 24 Sep 2021 09:13:32 +0000 (11:13 +0200)]
Release 5.4.8
Alexander Ebert [Fri, 24 Sep 2021 07:37:56 +0000 (09:37 +0200)]
Release 5.4.8 dev 2
Tim Düsterhus [Thu, 23 Sep 2021 11:33:54 +0000 (13:33 +0200)]
Merge pull request #4516 from WoltLab/xsrf-token-error
Improve phrasing for XSRF token error messages
Tim Düsterhus [Thu, 23 Sep 2021 10:47:15 +0000 (12:47 +0200)]
Improve phrasing in wcf.ajax.error.sessionExpired
see #4501
Tim Düsterhus [Thu, 23 Sep 2021 10:44:32 +0000 (12:44 +0200)]
Improve phrasing in wcf.global.form.error.securityToken
see #4501
Alexander Ebert [Wed, 22 Sep 2021 16:35:39 +0000 (18:35 +0200)]
Release 5.4.8 dev 1
WoltLab [Wed, 22 Sep 2021 16:11:07 +0000 (16:11 +0000)]
Updating minified JavaScript files
Tim Düsterhus [Wed, 22 Sep 2021 12:55:44 +0000 (14:55 +0200)]
Fix typo in setup_en.xml
Marcel Werk [Wed, 22 Sep 2021 08:57:09 +0000 (10:57 +0200)]
Incorrect type comparison when the legacy mysql extension is been used
Tim Düsterhus [Tue, 21 Sep 2021 14:58:07 +0000 (16:58 +0200)]
Merge branch '5.3' into 5.4
Marcel Werk [Tue, 21 Sep 2021 14:53:14 +0000 (16:53 +0200)]
Merge pull request #4497 from max-m/patch-categoryMultiSelectOptionType
Make `categoryMultiSelectOptionType.tpl` behave like `categoryOptionList.tpl`
Tim Düsterhus [Tue, 21 Sep 2021 14:31:17 +0000 (16:31 +0200)]
Take the array key into account when checking whether an unnamed KEY matches in DatabaseTableChangeProcessor
The reproducer effectively matches
d7f721d6f920d66f75102723b504d89e57a8c9ff, except that the KEY
is unnamed.
Previously the update would silently fail to do anything. Now the update fails
loudly, because it attempts to create another index with an existing name. This
is no different behavior compared to an INDEX collision of two unnamed indices
`(a, b)`, `(a, c)`. The developer will be clearly alerted of this issue and can
take appropriate measures to avoid it, e.g. by using explicit names.
see #4434
Alexander Ebert [Tue, 21 Sep 2021 13:23:24 +0000 (15:23 +0200)]
Skip desktop notifications on Android
Notifications are not supported outside of the context of service workers.
See https://community.woltlab.com/thread/292374-chrome-android-failed-to-construct-notification-illegal-constructor/
Tim Düsterhus [Tue, 21 Sep 2021 08:59:22 +0000 (10:59 +0200)]
Do not error during validation of TOTP codes if an invalid device is selected
Tim Düsterhus [Tue, 21 Sep 2021 08:54:46 +0000 (10:54 +0200)]
Do not pass `null` to `|encodeJS`
This breaks in PHP 8.1.
Tim Düsterhus [Mon, 20 Sep 2021 16:51:04 +0000 (18:51 +0200)]
Fix TypeScript code style
Tim Düsterhus [Mon, 20 Sep 2021 16:50:03 +0000 (18:50 +0200)]
Fix SCSS code style
Tim Düsterhus [Mon, 20 Sep 2021 16:47:42 +0000 (18:47 +0200)]
Use well-specified node.js for Prettier jobs in GitHub Actions
Tim Düsterhus [Mon, 20 Sep 2021 16:44:54 +0000 (18:44 +0200)]
Update GitHub Actions to node.js 16
Alexander Ebert [Mon, 20 Sep 2021 16:25:10 +0000 (18:25 +0200)]
Skip bogus selection changes
Alexander Ebert [Mon, 20 Sep 2021 16:20:28 +0000 (18:20 +0200)]
Skip the check for the caret position if the selection is invalid
Alexander Ebert [Mon, 20 Sep 2021 15:48:46 +0000 (17:48 +0200)]
Merge branch '5.3' into 5.4
Alexander Ebert [Mon, 20 Sep 2021 15:48:31 +0000 (17:48 +0200)]
Incorrect gradient value in Safari
https://community.woltlab.com/thread/292475-mainmenushowprevious-mainmenushownext-safari-farbunterschied-fehler/
Tim Düsterhus [Mon, 20 Sep 2021 15:15:31 +0000 (17:15 +0200)]
Update npm dependencies
Tim Düsterhus [Mon, 20 Sep 2021 14:52:58 +0000 (16:52 +0200)]
Fix SCSS code style
Apparently the replacement of 100% by 1 caused the line to be sufficiently
short to rewrap.
Tim Düsterhus [Mon, 20 Sep 2021 14:47:14 +0000 (16:47 +0200)]
Fix use of transparentize() in *.scss
The function expects a unitless number between 0 and 1 as the second parameter.
The updated SCSS compiler in 5.5+ complains about this misuse.
Tim Düsterhus [Mon, 20 Sep 2021 14:23:41 +0000 (16:23 +0200)]
Fix return type for SeekableIterator::* implementations
The `mixed` type is not available with our current minimum PHP version, thus
using ReturnTypeWillChange for these.
Tim Düsterhus [Mon, 20 Sep 2021 14:01:02 +0000 (16:01 +0200)]
Move scssphp to the WoltLab fork
This is for PHP 8.1 compatibility.
Alexander Ebert [Mon, 20 Sep 2021 13:58:59 +0000 (15:58 +0200)]
Merge branch '5.4' of https://github.com/WoltLab/WCF into 5.4
Alexander Ebert [Mon, 20 Sep 2021 13:58:51 +0000 (15:58 +0200)]
Enable `X-Frame-Options` for the WCFSetup
This has the side effect of suppressing `SameSite=none` for the cookies, which fails on insecure connections because this attribute value is valid for secure cookies only.
Resolves #4499
Follow up for
2a9d48c4badc4de2e0f2d2fc73c3af2bee39cce8
Tim Düsterhus [Mon, 20 Sep 2021 13:48:03 +0000 (15:48 +0200)]
Fix return type of `count()` for the remaining classes implementing \Countable
see
3f6b343d10b044ab08d41fec525f69ffe0a95e49
Tim Düsterhus [Mon, 20 Sep 2021 13:44:14 +0000 (15:44 +0200)]
Fix PHP 8.1 compatibility in DatabaseObjectList
> Return type of wcf\data\DatabaseObjectList::count() should either be
> compatible with Countable::count(): int, or the #[ReturnTypeWillChange]
> attribute should be used to temporarily suppress the notice […]
Alexander Ebert [Mon, 20 Sep 2021 13:31:54 +0000 (15:31 +0200)]
Enable `X-Frame-Options` for the WCFSetup
This has the side effect of suppressing `SameSite=none` for the cookies, which fails on insecure connections because this attribute value is valid for secure cookies only.
Resolves #4499
Alexander Ebert [Mon, 20 Sep 2021 12:17:50 +0000 (14:17 +0200)]
Skip the default cover photo when rebuilding users
Fixes #4500
Tim Düsterhus [Mon, 20 Sep 2021 11:37:54 +0000 (13:37 +0200)]
Set the XSRF-Token cookie to SameSite=lax
As it turns out, `strict` is too strict for some use cases of the average user,
as it might suppress the cookie when the user researches something while
writing a post and ultimately comes back to the community via an external link.
This request will not have the XSRF-Token cookie attached due to violating the
`strict` policy, resulting in WoltLab Suite sending a fresh cookie in response.
This will then invalidate the token stored in the form where the user is in the
process of writing their post, ultimately resulting in an error message.
The `SameSite` value is meant as a defense in depth measure to protect the user
even if they current token leaked in some way. Reducing the strictness does not
reduce the security in a measurable way.
Marcel Werk [Fri, 17 Sep 2021 12:15:30 +0000 (14:15 +0200)]
Typo
Maximilian Mader [Tue, 14 Sep 2021 16:19:15 +0000 (18:19 +0200)]
Make `categoryMultiSelectOptionType.tpl` behave like `categoryOptionList.tpl`
Currently the `categoryMultiSelectOptionType.tpl` outputs nodes of depth 0 and depth 1,
but article categories for example can be nested deeper than that.
The `categoryOptionList.tpl` as used by the category add forms handles arbitrary nesting levels already,
so I’ve ported the template logic over to the multi select option template.
The maximum nesting depth in the `AbstractCategoryMultiSelectOptionType.class.php` has been changed
to the default value (-1) to allow for infinite nesting depths.
Alexander Ebert [Sat, 11 Sep 2021 15:07:28 +0000 (17:07 +0200)]
Merge branch '5.3' into 5.4
Alexander Ebert [Sat, 11 Sep 2021 15:05:59 +0000 (17:05 +0200)]
Merge pull request #4496 from mutec/tagfieldfdp1
fix id of `CustomFormDataProcessor` in `TagFormField`
mutec [Fri, 10 Sep 2021 14:49:57 +0000 (16:49 +0200)]
fix id of `CustomFormDataProcessor` in `TagFormField`
the id was `acl` which seems to be a copy paste-mistake
Alexander Ebert [Thu, 9 Sep 2021 11:47:17 +0000 (13:47 +0200)]
Incorrect code style
Alexander Ebert [Thu, 9 Sep 2021 11:42:47 +0000 (13:42 +0200)]
Adjusted the `tab_width` for *.css/*.scss
Alexander Ebert [Thu, 9 Sep 2021 11:35:31 +0000 (13:35 +0200)]
Explicitly reduce the indent size of *.css/*.scss
Alexander Ebert [Thu, 9 Sep 2021 09:58:15 +0000 (11:58 +0200)]
Merge branch '5.3' into 5.4
Alexander Ebert [Thu, 9 Sep 2021 09:58:00 +0000 (11:58 +0200)]
Merge branch '5.2' into 5.3
Alexander Ebert [Thu, 9 Sep 2021 09:56:06 +0000 (11:56 +0200)]
Merge branch '5.3' into 5.4
Alexander Ebert [Thu, 9 Sep 2021 09:50:40 +0000 (11:50 +0200)]
Merge pull request #4495 from Fabii547/patch-79
Fix writing `runStandalone` attribute to package.xml
Alexander Ebert [Thu, 9 Sep 2021 09:47:02 +0000 (11:47 +0200)]
Merge pull request #4493 from WoltLab/5.3-aclformfieldcleanup
Reset ACL field values within form cleanup
Fabii547 [Thu, 9 Sep 2021 08:28:36 +0000 (10:28 +0200)]
Fix writing `runStandalone` attribute to package.xml
Alexander Ebert [Wed, 8 Sep 2021 13:38:46 +0000 (15:38 +0200)]
Release 5.4.7
Alexander Ebert [Wed, 8 Sep 2021 12:13:30 +0000 (14:13 +0200)]
Release 5.3.14
Alexander Ebert [Wed, 8 Sep 2021 12:05:13 +0000 (14:05 +0200)]
Incorrect use of spaces for indentation in <5.4
Alexander Ebert [Sun, 8 Aug 2021 09:29:26 +0000 (11:29 +0200)]
Sandbox `foreachVars` in templates
Nesting the same template inside a `foreach` loop that is also accessed inside the nested call will overwrite the values from the outer template due to identical identifiers being used.
The sandbox did not protected `$this->foreachVars` despite being stateful.
See #4431
Fixes #4444
joshuaruesweg [Wed, 8 Sep 2021 11:53:42 +0000 (13:53 +0200)]
Reset ACL field values within form cleanup
joshuaruesweg [Wed, 8 Sep 2021 11:51:29 +0000 (13:51 +0200)]
Remove empty lines
Alexander Ebert [Tue, 7 Sep 2021 21:01:22 +0000 (23:01 +0200)]
Release 5.4.7 RC 2
Alexander Ebert [Tue, 7 Sep 2021 21:00:29 +0000 (23:00 +0200)]
Replaced a hard-wired color value, Safari fix
Alexander Ebert [Tue, 7 Sep 2021 11:33:58 +0000 (13:33 +0200)]
Release 5.4.7 RC 1
Alexander Ebert [Tue, 7 Sep 2021 11:32:43 +0000 (13:32 +0200)]
Merge branch '5.4' of https://github.com/WoltLab/WCF into 5.4
WoltLab [Tue, 7 Sep 2021 11:30:44 +0000 (11:30 +0000)]
Updating minified JavaScript files
Alexander Ebert [Tue, 7 Sep 2021 11:28:49 +0000 (13:28 +0200)]
Removed the upgrade instructions for 5.3.*
This greatly reduces the size of the update packages. Upgrade from 5.3 will still work, but will upgrade to 5.4.6 only and then apply the remaining updates.
Alexander Ebert [Tue, 7 Sep 2021 11:21:01 +0000 (13:21 +0200)]
Check against the current menu item during editing only