-
Notifications
You must be signed in to change notification settings - Fork 198
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
tree-wide: Remove (almost) all remaining rojig bits #2842
Conversation
I initially started by doing a search for everything referencing This series uses |
67b5d08
to
673b320
Compare
OK actually I went to try removing |
6c639b7
to
4eef894
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM!
Needs a rebase.
4eef894
to
c784a86
Compare
Nothing uses this code now.
Prep for removing the APIs entirely.
Nothing calls these now.
Has never been used.
One thing I realized is we need to keep the `rojig:` bit in the treefile because we're (ab)using it in coreos-assembler for image naming. But we don't need the spec file generation bits, so that can go.
The only part left that we will need to keep ~forever is the treefile parsing `rojig:` because it's used by coreos-assembler. But all we need is to propagate it into the JSON treefile.
c784a86
to
7082822
Compare
Ah ok, fixed a logic error in the origin code removal. |
Conceptually `image.yaml` is the configuration for coreos-assembler specifically, distinct from `manifest` and `overlay` etc. which end up being fed into rpm-ostree. First, this is part of addressing coreos/rpm-ostree#2842 (comment) so that e.g. we can migrate the `name:` into this file instead of the rpm-ostree manifest. But I also want to hook other things off here, such as whether to use the new ostree-container format instead of our tar-of-archive-repo.
Conceptually `image.yaml` is the configuration for coreos-assembler specifically, distinct from `manifest` and `overlay` etc. which end up being fed into rpm-ostree. First, this is part of addressing coreos/rpm-ostree#2842 (comment) so that e.g. we can migrate the `name:` into this file instead of the rpm-ostree manifest. But I also want to hook other things off here, such as whether to use the new ostree-container format instead of our tar-of-archive-repo.
Conceptually `image.yaml` is the configuration for coreos-assembler specifically, distinct from `manifest` and `overlay` etc. which end up being fed into rpm-ostree. First, this is part of addressing coreos/rpm-ostree#2842 (comment) so that e.g. we can migrate the `name:` into this file instead of the rpm-ostree manifest. But I also want to hook other things off here, such as whether to use the new ostree-container format instead of our tar-of-archive-repo.
Ping on this one, any objections to landing it? Would like to clear some dead wood here for more construction. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Whoops sorry, missed your update here!
This mainly undoes coreos@2c270a6, and follows up on coreos#2842. Also serves as some cleanup to make way for the introduction of a new `container-image-reference` refspec type. We no longer have any use for `ostree://` or `rojig://`-style prefixes in the refspecs. There had been efforts to "canonicalize" refspecs to always include such prefixes into the code internally, but that hasn't gotten too far. Crucially, since such prefixes were never introduced to libostree, the refspecs in the origin file were never meant to contain such prefixes anyway. Following some discussion in coreos#2940 (comment), in the future we would disambiguate "refspec type" via the name of the key in the origin file instead.
This mainly undoes coreos@2c270a6, and follows up on coreos#2842. Also serves as some cleanup to make way for the introduction of a new `container-image-reference` refspec type. We no longer have any use for `ostree://` or `rojig://`-style prefixes in the refspecs. There had been efforts to "canonicalize" refspecs to always include such prefixes into the code internally, but that hasn't gotten too far. Crucially, since such prefixes were never introduced to libostree, the refspecs in the origin file were never meant to contain such prefixes anyway. Following some discussion in coreos#2940 (comment), in the future we would disambiguate "refspec type" via the name of the key in the origin file instead.
This mainly undoes coreos@2c270a6, and follows up on coreos#2842. Also serves as some cleanup to make way for the introduction of a new `container-image-reference` refspec type. We no longer have any use for `ostree://` or `rojig://`-style prefixes in the refspecs. There had been efforts to "canonicalize" refspecs to always include such prefixes into the code internally, but that hasn't gotten too far. Crucially, since such prefixes were never introduced to libostree, the refspecs in the origin file were never meant to contain such prefixes anyway. Following some discussion in coreos#2940 (comment), in the future we would disambiguate "refspec type" via the name of the key in the origin file instead.
This mainly undoes coreos@2c270a6, and follows up on coreos#2842. Also serves as some cleanup to make way for the introduction of a new `container-image-reference` refspec type. We no longer have any use for `ostree://` or `rojig://`-style prefixes in the refspecs. There had been efforts to "canonicalize" refspecs to always include such prefixes into the code internally, but that hasn't gotten too far. Crucially, since such prefixes were never introduced to libostree, the refspecs in the origin file were never meant to contain such prefixes anyway. Following some discussion in coreos#2940 (comment), in the future we would disambiguate "refspec type" via the name of the key in the origin file instead.
This mainly undoes 2c270a6, and follows up on #2842. Also serves as some cleanup to make way for the introduction of a new `container-image-reference` refspec type. We no longer have any use for `ostree://` or `rojig://`-style prefixes in the refspecs. There had been efforts to "canonicalize" refspecs to always include such prefixes into the code internally, but that hasn't gotten too far. Crucially, since such prefixes were never introduced to libostree, the refspecs in the origin file were never meant to contain such prefixes anyway. Following some discussion in #2940 (comment), in the future we would disambiguate "refspec type" via the name of the key in the origin file instead.
Conceptually `image.yaml` is the configuration for coreos-assembler specifically, distinct from `manifest` and `overlay` etc. which end up being fed into rpm-ostree. First, this is part of addressing coreos/rpm-ostree#2842 (comment) so that e.g. we can migrate the `name:` into this file instead of the rpm-ostree manifest. But I also want to hook other things off here, such as whether to use the new ostree-container format instead of our tar-of-archive-repo.
importer: Remove rojig bits
core: Remove rojig bits
Nothing uses this code now.
daemon/sysroot: Replace some rojig code with g_assert_not_reached()
Prep for removing the APIs entirely.
origin: Remove rojig bits
Nothing calls these now.
rust/origin: Remove rojig bits
Has never been used.
util: Remove rojig cache branch mapping functions
Not used.
rust/treefile: Remove rojig spec writing
One thing I realized is we need to keep the
rojig:
bitin the treefile because we're (ab)using it in coreos-assembler
for image naming.
But we don't need the spec file generation bits, so that can go.
tree-wide: Remove (almost) all remaining rojig bits
The only part left that we will need to keep ~forever is
the treefile parsing
rojig:
because it's used by coreos-assembler.But all we need is to propagate it into the JSON treefile.