---
type: change
description: |-
add change_description field to credential commits
This also involved changing how commit range change fingerprints are handled, so
that the description can be more variable, and using that flexibility to allow
the cli tool the ability to ask for the change description prior to creating a
credential commit. Credential commit creation via the cli tool has been further
improved in that it will check if it's been told to make a cred based on a
specific credential commit, and if so just append the new credential and
re-commit, speeding things up a bit for everyone.
fingerprint: AGWXqPM1Lxf3kvV/V3w65jNiUQyrQfemqfCcImSWjHJl
credentials:
- type: pgp_signature
pub_key_id: 95C46FA6A41148AC
body: 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
account: mediocregopher
---
type: change
description: |-
Remove Payload.MessageHead error return and simplify implementations
None of the new implementations of Payload.MessageHead can return an error, so
don't do it. The new implementations place more reliance on the author field of
the commit to describe the "who", so the commit heads have more room to describe
the "what".
fingerprint: AIf/6pSXd2ao5vKa8Ju0HAKWwPC2xN2RDQtoUfuDZcnc
credentials:
- type: pgp_signature
pub_key_id: 95C46FA6A41148AC
body: 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
account: mediocregopher
---
type: change
description: |-
Fix a bug which prevented force pushing to a previous commit in a branch
This bug was caused because VerifyCanSetBranchHEADTo did not properly handle the
case; what would happen is that the merge-base of the old HEAD and the new one
would be taken, and then the range of that merge-base result to the new commit
would be retrieved. But since the merge-base of the old HEAD and the new would
just be the new HEAD, it'd be taking the range from new HEAD to new HEAD; and
empty range.
This commit adds an ancestry check prior to the merge-base, to manually account
for this case, and calls verifyCommit directly if it comes about. It's not very
pretty, but it works.
fingerprint: AEYAvGa1GIfbmjh/gqu2l7qdhQtlJIGHhJd1GaZ5D87h
credentials:
- type: pgp_signature
pub_key_id: 95C46FA6A41148AC
body: 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
account: mediocregopher
---
type: change
description: |-
Fix a bug in typeobj when a type field's name is the same as one of its inner fields
This specifically came up with Comment (though it wasn't caught because an error
wasn't being caught, that's fixed here as well). Prior to unmarshaling into the
selected inner struct field, typeobj.UnmarshalYAML unmarshals into the outer
struct in order to unmarshal all non-type fields. However, if one of the fields
intended for the inner struct field has the same name as one of the type fields
in the outer struct there would be a conflict at this point.
The solution is to modify the type of the outer struct being unmarshaled into at
this stage, so that all fields with type tags automatically have a `yaml:"-"`
tag, and so are ignored by the yaml unmarshaler at this stage.
fingerprint: AL32FBVJ7Bu2dz1ysrCiRFz2/y+QuaEyhKygvWP/fihw
credentials:
- type: pgp_signature
pub_key_id: 95C46FA6A41148AC
body: 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
account: mediocregopher
---
type: change
description: |-
Completely refactor naming of everything, in light of new SPEC
Writing the SPEC shed some light on just how weakly a lot of concepts, like
"commit", had been defined, and prompted the delineation of a lot of things
along specific lines (commit vs payload, repo vs project). This commit makes the
code reflect the SPEC much better in quite a few ways:
* Repo is now Project
* Commit is now Payload
* GitCommit is now just Commit
* Hash is now Fingerprint
* A lot of minor fields got renamed
* All the XXXInterface types are now just XXX, and their old XXX type is now
XXXUnion.
More than likely there's still some comments and variable names that have
slipped passed, but overall I feel like I got most of the changes.
fingerprint: AKkDC5BKhKbfXzZQ/F4KquHeMgVvcNxgLmkZFz/nP/tY
credentials:
- type: pgp_signature
pub_key_id: 95C46FA6A41148AC
body: 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
account: mediocregopher
