The keying modules tries to solve two problems, the lack of key
separation and the lack of AEAD being used for encryption. The currently
used `secrets` doesn't provide this and is hard to adjust to provide
this functionality.
For encryption, the additional data is now a parameter that can be used,
as the underlying primitive is an AEAD constructions. This allows for
context binding to happen and can be seen as defense-in-depth; it
ensures that if a value X is encrypted for context Y (e.g. ID=3,
Column="private_key") it will only decrypt if that context Y is also
given in the Decrypt function. This makes confused deputy attack harder
to exploit.[^1]
For key separation, HKDF is used to derives subkeys from some IKM, which
is the value of the `[service].SECRET_KEY` config setting. The context
for subkeys are hardcoded, any variable should be shuffled into the the
additional data parameter when encrypting.
[^1]: This is still possible, because the used AEAD construction is not
key-comitting. For Forgejo's current use-case this risk is negligible,
because the subkeys aren't known to a malicious user (which is required
for such attack), unless they also have access to the IKM (at which
point you can assume the whole system is compromised). See
https://scottarc.blog/2022/10/17/lucid-multi-key-deputies-require-commitment/
UX/Translation changes:
- new teams: remove redundant tooltips that don't add meaningful information
- move general information to table fieldset
- new teams: rename "general" to "custom" access for clarity
- new teams: show labels beside options on mobile
Accessibility:
- semantic form elements allow easier navigation (fieldset, mostly)
- improve better labelling of new teams table
- fix accessibility scan issues
- TODO: the parts that "disable" form elements were not yet touched and
are not really accessible to screenreaders
Technical:
- replace two JavaScript solutions with one CSS standard
- implement a simpler grid (.simple-grid)
- simplify markup
- remove some webhook settings specific CSS
Testing:
- check more form content for accessibility issues
- but exclude tooltips from the scan :(
- reuse existing form tests from previous PR
using middleware validator to validate title length on update
use error name from binding package
add integration test for title update
rebase upstream and update test var name
fix test slice formatting
just a try (#1)
Reviewed-on: https://codeberg.org/thilinajayanath/forgejo/pulls/1
Co-authored-by: Otto Richter <git@otto.splvs.net>
Co-committed-by: Otto Richter <git@otto.splvs.net>
fix errors + add test for 255 char title
fix test domain
fix CSRF token error on test
updaate result struct that's used to decode the json response
add json tags for struct and check changed title when http 200 is received
try to decode the title if the request succeeded
add comment in integration test
LDFLAGS="-buildid=" must be set in the environment so the Makefile
adds to it. Setting it via the make arguments overrides it and removes
the -X "main.*Version" arguments which are used to set the Forgejo
version of the binary.
Regression introduced in [CHORE] Support reproducible builds' (#4970)
Remove the snapping of the images on the projects cards, the images are
way too small to notice that when scrolling you're being snapped to
these images and when you do notice it, it doesn't make sense as you
wouldn't expect it to be snapped.
This is a step towards making Forgejo's binaries (the one listed in the
release tab) reproducible.
In order to make the actual binary reproducible, we have to ensure that
the release workflow has the correct configuration to produce such
reproducible binaries. The release workflow currently uses the
Dockerfile to produce binaries, as this is one of the easiest ways to do
cross-compiling for Go binaries with CGO enabled (due to SQLite). In the
Dockerfile, two new arguments are being given to the build command.
`-trimpath` ensures that the workpath directory doesn't get included in
the binary; this means that file names (such as for panics) are
relative (to the workpath) and not absolute, which shouldn't impact
debugging. `-buildid=` is added to the linker flag; it sets the BuildID
of the Go linker to be empty; the `-buildid` hashes the input actions
and output content; these vary from build to build for unknown reasons,
but likely because of the involvement of temporary file names, this
doesn't have any effect on the behavior of the resulting binary.
The Makefile receives a new command, `reproduce-build#$VERSION` which
can be used by people to produce a reproducible Forgejo binary of a
particular release; it roughly does what the release workflow also does.
Build the Dockerfile and extract the Forgejo binary from it. This
doesn't allow to produce a reproducible version for every release, only
for those that include this patch, as it needs to call the makefile of
that version in order to make a reproducible binary.
There's one thing left to do: the Dockerfile pins the Go version to a
minor level and not to a patch level. This means that if a new Go patch
version is released, that will be used instead and will result in a
different binary that isn't bit to bit the same as the one that Forgejo
has released.
One method to set them all... or something like that.
The defaults for git-grep options were scattered over the run
function body. This change refactors them into a separate method.
The application of defaults is checked implicitly by existing
tests and linters, and the new approach makes it very easy
to inspect the desired defaults are set.
