If a new branch is pushed in Git without new commits (the branch
is created on a commit that already exists in the SCM-Manager repository
and pushed without any further commit), all ancestors of the commit
the branch points to have been treated as if they were new by the
hook changeset provider. This led to severe errors like wrong
push logs (by the pushlog plugin) or re-evaluated commit messages
by the commit message checker plugin.
This fixes this wrong behaviour. If new commits are not provided by
the pack parser, no commit will be treated as a new one.
There is currently no documentation which shortcuts are available to the end user, neither within the application nor the documentation published on scm-manager.org. This PR adds the missing documentation in both places and introduces a new api for developers to add documentation when using `useShortcut`. It also improves the api for conditional shortcuts significantly.
If a POST request is submitted to the rest api for repostory permissions, the regex validator ignores the name field if it is null, which leads to an internal server error and breaks any further attempts to interact with that repository. An additional not-null constraint resolves this problem.
The global navigation can be accessed with the following key combination:
Option-R = Repositories
Option-U = User
Option-G = Groups
Option-A = Administration
The navigation within a repository can be accessed with a combination with g:
gi = Information
gb = Branches
gt = Tags
gc = Code
gs = Settings
Enable users to jump to the global search bar by pressing the "/" key from anywhere. Open modals block this shortcut. This PR also introduces a generic system for declaring global shortcuts by utilizing the third-party library mousetrap.
This reduces the amount of lines logged with log level
'trace' from the DefaultRealm. This is done by concatenating
the single permissions and roles in a single line.
A log with the new layout looks something like this (the [...]
would list all further permissions in the real log output):
TRACE sonia.scm.security.DefaultRealm - authorization summary:
username : scmadmin
roles : user
scope :
permissions: repository:*:ChTG0dBeUH , group:autocomplete , repository:*:[...]
In contrast, with the old layout it looks like this:
TRACE sonia.scm.security.DefaultRealm - authorization summary:
username : scmadmin
roles :
- user
scope :
permissions:
- repository:*:ChTG0dBeUH
- group:autocomplete
- repository:*:5DTFWwBTiY
- repository:*:ASTGBIVz11D
- user:changeApiKeys:scmadmin
- user:changePassword:scmadmin
- user:changePublicKeys:scmadmin
- user:readAuthorizedKeys,writeAuthorizedKeys:scmadmin
- *
- user:read:scmadmin
- user:autocomplete
- repository:*:4rT7VPex5J
The lazy loading feature implemented for git was broken,
because the repeated usage in the BrowserResultCollapser
has overwritten the request in the command. Therefore
the command could no longer update the cache in the
BrowseCommandBuilder.
To fix this, we now use a browse command factory (represented
by a simple supplier) that will create a dedicated command
implementation for each request issued in the collapser.
Add a preset for the depcheck library that is applied to all projects in scm-ui. Some minor dependency issues were discovered and resolved along the way. Some features of depcheck did not work correctly or were missing, which is why the configuration file is a bit more verbose than desired.
Using plus symbol in global search led to a space in the query parameter of the URL. We now encode the query string properly in global search and additionally fixed the expert search documentation.
Due to unexpected and largely unchangeable behavior by both react-router and the browser, square brackets are not correctly encoded in the url when clicking a file link in the source view where the filename contains either of these characters. The source view then tries to use the useSources hook to get the file content but fails, because the path param for the file path it gets from the url has unencoded square brackets in them which are illegal in urls except for declaring IPv6 addresses. We have created a catch for exactly this scenario at the latest possible point before the actual http request is fired, which is in the useSources hook. It seems like the square brackets are the only affected special characters so we force encoding on them specifically. Only the path portion of the URL is checked so the host portion of the url may still contain unencoded square brackets which are left untouched.
We stumbled upon errors accessing index from plugins which were not yet initialized. To prevent this errors we use our own No-Op IndexReader for missing indices.
The new docker health check respects the jetty configuration and uses the configured port and context path.
It supports ssl listeners and follows redirects.
Co-authored-by: Sebastian Sdorra <sebastian.sdorra@cloudogu.com>
After many days invested in making tailwind work in the SCM-Manager environment as well as a long discussion last week, we have decided not to move further with tailwind, but still keep adding new, independent modules for frontend components. Tailwind simply overcomplicated our build pipeline because bulma was already part of the api and the two were incompatible on several occasions. Styling will continue to be guided by bulma and all parts related to tailwind are removed. We therefore continue the trend of focusing on improving our existing stack rather than adding further complexity.
Enhance search result view by sorting the categories after their translation (repositories will still be sticky on top). Further disable categories with no search results and be more explicit with the text displayed if no search results were found.
Adds a new button to repository settings to allow users to manually delete and re-create search indices. The actual re-indexing is happening in plugins that subscribe to the newly created event.
Co-authored-by: Eduard Heimbuch <eduard.heimbuch@cloudogu.com>
