* Don't manually replace whitespace during render
For historical reasons Gitea manually alters the urlPrefix and replaces
a whitespace with a +. This Works for URLs, but we're also passing
urlPrefix to git calls and adding the + is breaking the tree path.
Goldmark will automatically convert a white space to the proper %20, so
we should leave the string as is which lets us pass it to git unmodified
and then let Goldmark fix it.
Also fixed separate bug in URLJoin I noticed while testing where it will
silently discard sections of a path that have # in them (possibly
others). We should just escape it first.
Fixes 10156
* Escape elems as well
* Revert "Escape elems as well"
This reverts commit 8bf49596fe.
* restart ci
* remove changes to URLJoin
* restart ci
Co-authored-by: techknowlogick <matti@mdranta.net>
Co-authored-by: techknowlogick <matti@mdranta.net>
* fix pull view when head repository or head branch missed and close related pull requests when delete branch
* fix pull view broken when head repository deleted
* close pull requests when head repositories deleted
* Add tests for broken pull request head repository or branch
* fix typo
* ignore special error when close pull request
Co-authored-by: Lauris BH <lauris@nix.lv>
Co-authored-by: Lauris BH <lauris@nix.lv>
* Fix wrong hint when status checking is running on pull request view
* fix lint
* fix test
* fix test
* fix wrong tmpl
* fix import
* rename function name
* Fix wrong permissions check when issues/prs shared operations
* move redirect to the last of the function
* fix swagger
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: Lauris BH <lauris@nix.lv>
* ci: use docker image for golangci-lint (#9737)
* ci: re-ordering Drone CI for optimizing time (#9719)
* ci: try re-ordering for optimizing time
* ci: try re-ordering for optimizing time
* ci: try re-ordering for optimizing time
* ci: try re-ordering for optimizing time
* ci: try re-ordering for optimizing time
* ci: try re-ordering for optimizing time
* ci: try offloading mysql8 to arm64
* Revert "ci: try offloading mysql8 to arm64"
This reverts commit c60de5db1c.
* ci: try offloading pgsql to arm64
* ci: activate ldap on arm64
* ci: test mysql8 in place pgsql arm64
* chore: clean un-needed move
* typo
* ci: revert runnning mysql on arm64
* ci: run compliance on arm
* chore: limit change
* chore: readd maybe need for release fetch-tags
* ci: remove docker-linux-amd64-dry-run
* ci: remove docker-linux-amd64-dry-run
* Revert "ci: remove docker-linux-amd64-dry-run"
This reverts commit 0715f65b11.
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
* ci: use new mssql image (#9720)
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
* ci: run notify on arm64 (#9762)
* ci: run notify on arm64
Free one jobs on amrd64
* Update .drone.yml
* Update .drone.yml
* Update .drone.yml
Based on: https://github.com/appleboy/drone-discord/blob/master/.drone.yml#L339
* improve trigger
Co-authored-by: techknowlogick <matti@mdranta.net>
* ci: move some integration tests on arm64 (#9747)
* tests: configure github remaining limit + read token (#9800)
* ci: configure remaining github limmit
* prepend with github since package is common to all migrations
* add RefreshRate
* Update github.go
* add missing space
* go fmt
* Read env variable GITHUB_READ_TOKEN for token
* Update .drone.yml
* ci: simplify tag/release by always running coverage (#9774)
* ci: simplify tag/release by always running coverage
* use mod and vendor for unit test coverage
* remove not needed lfs for unit test
* use arm drone agent for docs (#9776)
* run translations pipeline on arm server (#9865)
* add git-check to Makefile
Co-authored-by: Antoine GIRARD <sapk@users.noreply.github.com>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
Co-authored-by: techknowlogick <matti@mdranta.net>
* Make CertFile and KeyFile relative to CustomPath
The current code will absolute CertFile and KeyFile against the current working directory. This is quite unexpected for users. This code makes relative paths absolute against the CustomPath.
Fix#4196
* Improve error reporting when reading certificates
* Apply suggestions from code review
Co-Authored-By: guillep2k <18600385+guillep2k@users.noreply.github.com>
Co-authored-by: guillep2k <18600385+guillep2k@users.noreply.github.com>
Co-authored-by: guillep2k <18600385+guillep2k@users.noreply.github.com>
* Fix missing updated time on migrated issues and comments
* Fix testing and missing updated on migrating pullrequest
Co-authored-by: Antoine GIRARD <sapk@users.noreply.github.com>
Co-authored-by: Antoine GIRARD <sapk@users.noreply.github.com>
* Delay printing hook statuses until after 1 second
* Move to a 5s delay, wrapped writer structure and add config
* Update cmd/hook.go
* Apply suggestions from code review
* Update cmd/hook.go
Co-authored-by: Antoine GIRARD <sapk@users.noreply.github.com>
Co-authored-by: Antoine GIRARD <sapk@users.noreply.github.com>
* Some more e-mail notification fixes
A few more small e-mail notification fixes/changes
* Style footer of notification email to be smaller
* Include text for when pull request is merged
* Don't include original body of issue or PR when merging/closing by
setting issue.Content to "" in these cases
* Set Re: prefix and meessage-ID headers based on actName instead of checking for a
comment. This fixes a bug where certain actions that didn't have a
comment were setting Message-ID instead of In-Reply-To which caused some
mail programs not to show those messages as they would have had the same
Message-ID as a previous message. Also fixes the case where a final
comment and closing message would have been displayed out of order if
you didn't have a copy of the original issue/pr cretion message.
* Update other template footers for consistency
* Move to goldmark
Markdown rendering moved from blackfriday to the goldmark.
Multiple subtle changes required to the goldmark extensions to keep
current rendering and defaults.
Can go further with goldmark linkify and have this work within markdown
rendering making the link processor unnecessary.
Need to think about how to go about allowing extensions - at present it
seems that these would be hard to do without recompilation.
* linter fixes
Co-authored-by: Lauris BH <lauris@nix.lv>
Instead of just adding test generated files to .gitignore prevent
them from being produced in the first place.
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
* make notifyWatchers work on multiple actions
* more efficient multiple notifyWatchers
* Make CommitRepoAction take advantage of multiple actions
* Batch post and pre-receive results
* Set batch to 30
* Auto adjust timeout & add logging
* adjust processing message
* Add some messages to pre-receive
* Make any non-200 status code from pre-receive an error
* Add missing hookPrintResults
* Remove shortcut for single action
* mistaken merge fix
* oops
* Move master branch to the front
* If repo was empty and the master branch is pushed ensure that that is set as the default branch
* fixup
* fixup
* Missed HookOptions in setdefaultbranch
* Batch PushUpdateAddTag and PushUpdateDelTag
Co-authored-by: Lunny Xiao <xiaolunwen@gmail.com>
* Wrap the code indexer
In order to prevent a data race in the code indexer it must be wrapped
with a holder otherwise it is possible to Search/Index on an
incompletely initialised indexer, and search will fail with a nil
pointer until the repository indexer is initialised.
Further a completely initialised repository indexer should not be closed
until Termination otherwise actions in Hammer/Shutdown phases could
block or be lost.
Finally, there is a complex dance of shutdown etiquette should the index
initialisation fail. This PR restores that.
* Always return err if closed whilst waiting
Co-authored-by: techknowlogick <matti@mdranta.net>
