Previously, if a user was a guest member of a private project, they
could access the merge request template as we were not checking
permission-levels of the user.
When a issue template is asked for, the user must have :read_issue for
the project; or :read_merge_request when a merge request template is
asked for.
We also now rescue_from FileNotFoundError and handle as 404. This is
because RepoTemplateFinder can raise a FileNotFoundError exception,
which Rails previously handled as a 500.
Handling these in a way that is consistent with
ActiveRecord::RecordNotFound exceptions, within controllers that
inherit from Projects::ApplicationController at least, and returning a
404.
https://gitlab.com/gitlab-org/gitlab-ce/issues/54943
By visiting `projects/:id` you will be redirected to project page with
path in it.
projects/123 => foo/bar
Signed-off-by: Dmitriy Zaporozhets <dmitriy.zaporozhets@gmail.com>
Repository archives are always named `<project>-<ref>-<sha>` even if
the ref is a commit. A consequence of always including the sha even
for tags is that packaging a release is more difficult because both
the ref and sha must be known by the packager.
- add `<project>/-/archive/<ref>/<filename>.<format>` route using the
`-` separator to prevent namespace collisions. If the filename is
`<project>-<ref>` or the ref is a sha, the sha will be omitted,
otherwise the default filename will be used.
- deprecate previous archive route `repository/<ref>/archive`
* implemented logging of project and system web hooks
* implemented UI for user area (project hooks)
* implemented UI for admin area (system hooks)
* implemented retry of logged webhook
* NOT imeplemented log remover
- gl_emoji for falling back to image/css-sprite when the browser
doesn't support an emoji
- Markdown rendering (Banzai filter)
- Autocomplete
- Award emoji menu
- Perceived perf
- Immediate response because we now build client-side
- Update `digests.json` generation in gemojione rake task to be more
useful and include `unicodeVersion`
MR: !9437
See issues
- #26371
- #27250
- #22474
==================
= Implementation =
==================
1. The path of the page is of the form 'group/project/pages/domains/<domain_name>'
2. Rails looks at `params[:id]` (which should be the domain name), and finds the
relevant model record.
3. Given a domain like `foo.bar`, Rails sets `params[:id]` to `foo` (should be
`foo.bar`), and sets `params[:format]` to `bar`
4. This commit fixes the issue by adding a route constraint, so that
`params[:id]` is set to the entire `foo.bar` domain name.
=========
= Tests =
=========
1. Add controller specs for the `PagesDomainController`. These are
slightly orthogonal to this bug fix (they don't fail when this bug is
present), but should be present nonetheless.
2. Add routing specs that catch this bug (by asserting that the `id`
param is passed as expected when it contains a domain name).
3. Modify the 'RESTful project resources' routing spec shared example to
accomodate controllers where the controller path (such as
`pages/domains`) is different from the controller name (such as
`pages_domains`).
Finished up autocomplete_sources action and added frontend to fetch data only when its needed
Added wait_for_ajax to specs
Fixed builds and improved the setup/destroy lifecycle
Changed global namespace and DRYed up loading logic
Added safety for accidentally loading data twice
Removed destroy as its not necessary and is messing with click events from a blur race condition
Created AutocompleteSourcesController and updated routes
Fixed @undefined from tabbing before load ends
Disable tabSelectsMatch until we have loaded data
Review changes
We need to do two things to support this:
1. Simplify the regex capture in the routing for the CommitsController
to not exclude the '.atom' suffix. That's a perfectly valid git
branch name, so we shouldn't blow up if we get it.
2. Because Rails now can't automatically detect the request format, add
some code to do so in `ExtractPath` when there is no path. This means
that, given branches 'foo' and 'foo.atom', the Atom feed for the
former is unroutable. To fix this: don't do that! Give the branches
different names!
Fix problem with projects ending with .keys #3076Closes#3076
Move route `:username.keys` below project's routes. This allow project's to handle names ending with `.keys`
See merge request !1883
Used mime-types gem instead of hardcoding content types.
Allow multiple extensions in archive route (.tar.gz, .tar.bz2).
Change content disposition from infile(?) to attachment for api.
Fixed api would return “archive” instead of {project}-{hash}.{ext}