Backport #15015
Unfortunately there is a subtle problem with recreatetable on postgres which
leads to the sequences not being renamed and not being left at 0.
Fix#14725
Signed-off-by: Andrew Thornton <art27@cantab.net>
Backport #14986
* Fix an issue with panics related to attributes
* Wrap goldmark render in a recovery function
* Reduce memory use in render emoji
* Use a pipe for rendering goldmark - still needs more work and a limiter
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: Lauris BH <lauris@nix.lv>
* API: fix set milestone on PR creation
pr creation via API failed with 404, because we searched
for milestoneID 0, due to uninitialized var usage D:
* add tests
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: 6543 <6543@obermui.de>
Backport #14979
Due a problem with the ast.Walker in the our transformer in goldmark
an image with a sibling image will not be transformed to gain a parent
link. This PR fixes this.
Fix#12925
Signed-off-by: Andrew Thornton <art27@cantab.net>
When editing forked repos using the API the BaseRepository needs to loaded
in order to check its visibility otherwise there will be NPE panic.
Fix#14956
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: zeripath <art27@cantab.net>
Backport #14925
PR #13610 unfortunately disabled importing repositories from local paths.
This PR restores this functionality.
Fix#14700
Signed-off-by: Andrew Thornton <art27@cantab.net>
Backport #14895
Continuing on from #14888
The previous implementation has race whereby an incomplete upload or
hash mismatch upload can end up in the ContentStore. This PR moves the
validation into the reader so that if there is a hash error or size
mismatch the reader will return with an error instead of an io.EOF
causing the storage to abort the storage.
Signed-off-by: Andrew Thornton <art27@cantab.net>
LocalStorage should only put completed files in position
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: zeripath <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
Backport #14831
Unfortunately `git log revision ... --skip=x -- path` skips the number of commits
not the number of commits relating to the path.
This PR changes the function to have a reader that reads and skips the
necessary number of commits by hand instead.
Fix#8716
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
Co-authored-by: 6543 <6543@obermui.de>
Backport #14823
When Deleting Repository only explicitly close PRs whose base is not this repository
Fix#14775
Signed-off-by: Andrew Thornton <art27@cantab.net>
Backport #14834
When using HCaptcha on link account pages the site key needs to be passed
in. This PR ensures that HCaptchaSiteKey is set in the data.
Fix#14766
Signed-off-by: Andrew Thornton <art27@cantab.net>
Backport #14804
* CutDiffAroundLine makes the incorrect assumption that `---` and `+++` always represent part of the header of a diff.
This PR adds a flag to its parsing to prevent this problem and adds a streaming parsing technique to CutDiffAroundLine using an io.pipe instead of just sending data to an unbounded buffer.
Fix#14711
* Handle unquoted comment patch files
When making comment patches unfortunately the patch does not always quote the filename
This makes the diff --git header ambiguous again.
This PR finally adds handling for ambiguity in to parse patch
Fix#14812
* Add in testing for no error
There is no way currently for CutDiffAroundLine in this test to cause an
error however, it should still be tested.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Backport #14802
Instead of causing a log.Fatal, we should handle broken OAuth2
providers by disabling them.
Fix#8930
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: techknowlogick <techknowlogick@gitea.io>
Backport #14678
Gitea runs diff on highlighted code fragment for each line in order to
provide code highlight diffs. Unfortunately this diff algorithm is not
aware that span tags and entities are atomic and cannot be split.
The current fixup code makes some attempt to fix these broken tags
however, it cannot handle situations where a tag is split over multiple
blocks.
This PR provides a more algorithmic fixup mechanism whereby spans and
entities are completely coalesced into their respective blocks.
This may result in a incompletely reduced diff but - it will definitely
prevent the broken entities and spans that are currently possible.
As a result of this fixup several inconsistencies were discovered in our
testcases and these were also fixed.
Fix#14231
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
This PR improves HasPreviousCommit to prevent the automatic and recursive loading
of previous commits using git merge-base --is-ancestor and git rev-list
Fix#13684
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
Backport #14651
There is potentially a race with a slow starting internal
queue causing a NPE if Has is checked before the internal
queue has been setup.
This PR adds a lock on the Has() fn.
Fix#14311
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: 6543 <6543@obermui.de>
* Turn default hash password algorightm back to pbkdf2 from argon2 until we found a better one
* Add a warning on document
Co-authored-by: zeripath <art27@cantab.net>
Backport #14624
GetCommit() assumes that all 40 char strings are SHA1s. This leads to an
error if you try to do a PR on a branch which is 40 characters long.
This PR attempts the SHA first - and if it fails will switch to using rev-parse.
Fix#14470
Signed-off-by: Andrew Thornton <art27@cantab.net>
Per #14531, deleting a user account will delete the user's GPG keys
from the `gpg_key` table but not from `gpg_key_import`, which causes
an error when creating an account with the same email and attempting
to re-add the same key. This commit deletes all entries from
`gpg_key_import` that match any GPG key IDs belonging to the user.
Co-authored-by: Anton Khimich <anton.khimicha@mail.utoronto.ca>
Migrations currently uses the default Xorm mapper which is
not the same as the mapper Gitea actually uses.
This means that there is a difference between the struct
parsing and mapping to database tables in migrations as
compared to normal Sync2.
This was the cause for the catastrophic problem in v168 -
untagged fields are not mapped in the same way in migrations
as compared to outside of migrations.
This is also likely the cause of some weird subtle failures
in other migrations as any untagged field may not be being
mapped exactly the same way.
This PR suggests that we ensure that the mapper is set at
the start of the migrations code - but also enforces a strict
clean mapper between each migration.
Signed-off-by: Andrew Thornton <art27@cantab.net>
Co-authored-by: zeripath <art27@cantab.net>
6543