| Age | Commit message (Collapse) | Author |
|---|
|
but behind local.
correctly report on situation and allow push
|
|
instead of the full url so messages are easier to read
|
|
so it is clearer and easier to read
|
|
so it is clearer and easier to read
|
|
1. don't attempt to push to a remote which is already up-to-date
2. don't attempt to delete branch on remote if it is already deleted
3. only push when out of sync if remote tip is ancestor of pushed
commit
4. force push to remote if user force pushed and remote is in sync
with nostr
|
|
In preparation for its reuse within `push`
|
|
so we have a status update and know which server is slow
|
|
so it is more concise and clear
|
|
with git servers.
this will also form the basis of preventing failed pushes
|
|
as it was hardcoded to test value
|
|
This reverts commit 3d2e031294b467be725a4ac5f5e65e029d25699a.
|
|
use stderr to write lines that get displayed when run
as a git-remote-helper
|
|
This reverts commit 6b63f150ba9e7c69bf83bcbd24543051c765c9f6.
|
|
use `stderr` to write lines that get displayed when run
as a git-remote-helper
|
|
as `stderr` is being used when it shouldn't be
|
|
This reverts commit 4a38d5f4179114fce60f84c4694284a49f648c64.
|
|
unless there were errors and detailed report doesnt clear
|
|
as rust-nostr doesn't add trailing slash to relays
|
|
instead of manually set. aligns with ngit version.
|
|
support with optional relays as query parameter
|
|
in prep for supporting new format
|
|
listed in the announcement `clone` tag
|
|
initially we push to the first server, if successful update the
state event, then siliently push to the others and silently fail
|
|
so that if the git server is ahead of the state announcement,
only the commits in the state annoucement will be downloaded
|
|
and falls back to git server is state event cant be found
|
|
ensure refs are included in state event
use `HashMap` to improve `RepoState` struct
|
|
if no previous state events can be found it replicates git server
otherwise it just updates pushed value
|
|
to more easily access state details
|
|
and move code identifying git server url into `list`, `fetch` and `push`.
|
|
in preparation for managing state via nostr
|
|
if multiple maintainers produce state events with the same
timestamp, the event with the largest event_id will be used
|
|
only update remote ref when push was successful
|
|
so main doesn't grow too large
|
|
pass outcome back to git to enable reporting
|
|
requests so that the git_server is called once
rather than many time in serial
|
|
by using GitAuthenticatior credential callback directly
rather than its `push` method
|
|
push was not updating remote refs after the anon remote was being pushed
|
|
in the future we could implement our own credentials callback function
to remove the dependancy
|
|
as a simple proxy to the first git server listed in announcement
parse clone url as `nostr://naddr123...`
|
|
add missing or updated identifier to maintainers.yaml
as we were not checking whether it has changed or added
also update if relays do not match
|
|
instead of Kind::Custom(u16) as v33 of rust-nostr introduced them
|
|
if branch is ahead of main or master
|
|
to prevent accidental name conflicts. also moved to prs/* namespace
`pull` and `push` integration tests are intermitantly failing to end
at least for `push` they work when run individually but not
when run together
|
|
when finding repo announcements
add a hint to find naddr on gitworkshop.dev
|
|
bump all rust-nostr packages
move from using patched nip46 nip05 function to profile function in v33
fix send_events_to error message
|
|
when looking for proposal roots 'pull' was not filtering out
proposal revisions
issue identified in failing test, which it now passes
|
|
as part of a project to use fetch and the stored cache everywhere
|
|
as soon as they are successfully sent to at least one relay
|
|
as part of a project to use fetch and the stored cache everywhere
|
|
the last set of pull integration test fails:
when_latest_event_rebases_branch
we are planning on replacing pull so I'm not sure
whether it is worth fixing
|
|
as part of a project to use `fetch` and the stored cache everywhere
|
|
to proposal which is displayed at the end of the functon
|
|
reworking the tests and test suite as appropriate
|
|
when fetching a user profile from user relays it throw an error
when an existing version of the profile wasn't found
|
|
prevent an infinite loop when a new maintainer is identified
that hasn't issued a repo event yet
make it clear when repo_coordinates must not have relays
|
|
as any profile event was being recorded as a update
|
|
rather than ending the loop without fetching them
|
|
prevent infinite loop when new coordinates are found
|
|
into cache directory rather than config directories
also removed CodeCollaboration from path as this isn't helpful
whilst this is a breaking change, the cache was only introduced
during development of this version so it is not
highlighted as such
|
|
rather than as new profiles
|
|
add a message for the temporary situation where a maintainers.yaml
exists without an identifier listed
|
|
add the ability to fetch more than just the current user from
their user write relays
|
|
which will enable the following`list` test to pass once `fetch`
is intergrated into `list`:
finds_based_on_naddr_on_embeded_relay
|
|
only the profile from the saved user was being fetched.
tests are using cli login parameters and expecting to see the
user's name but it was only showing the npub.
fixed by allowing the explicit request of specfic user profiles.
|
|
set repo.nostr to naddr reflecting the announcement just issued
|
|
fetch automatically gets updates to logged in user profile / relays
fetching without specifying repo pointers will just fetch user
profiles so that can be used during login, if user profile isn't
in cache
login now uses fetch
|
|
just from repository / fallback relays and only if we don't already
have a version in cache
this is because we are only using it for the user's name
|
|
as 'fetching... updates:' feels more like further updates are
being fetched than 'fetching... found'
|
|
so it only counts commits and statues to existing proposals
|
|
to increase the likelihood that all events are found
|
|
im preparation for identifying new inbox relays
|
|
enabler to add simplicity, efficency and offline
capability to other functions
improve repo announcement selection
|
|
show open proposal by defult and include options to filter
by other statuses
there are not tests for this currently as the intention is
to transform 'list' further by adding a 'fetch'
|
|
a typo prevented fetching of saved bunker-uri and therefore ngit
would act as if it wasn't saved
|
|
disable event limits in rust-nostr which cause large patches to be filtered out
|
|
if the maintainers tag is omitted
|
|
instead of just tagging the first maintainer's repo event and each
maintainer with a p tag
This allows for easier discoverability of the proposal when:
* the first maintainer hasn't issued a repo event
* the maintainers change over time and the single tagged repo event
is no listed as a maintainer in anyone elses repo event
|
|
currently using patched version of rust-nostr with function to fetch
nip46 relays from nip05 providers. this patch has been merged so it
will make it into the next rust-nostr release.
|
|
and save details in git config
|
|
so that nip46 bunker signing can be added
|
|
ensure fresh global cache when under test conditions by conditionally
storing it in local ./git folder
|
|
replace ngit yaml file config with:
* nsec / ncryptsec / npub in git config in nostr.* namespace
* sql database cache for metadata and relay events
allow different logins to be used for different git repositories by
storing login in local git config
|
|
local or global but tests only added for local
|
|
based on nip34 update.
see nip repository commit 8fe6e062254b37f77540088cccff60fa8615751
|
|
both nostr and nostr-sdk packages and also in test_utils
fix the many breaking changes
fix: ignore trailing slash when depuplicate relays for send events.
this was picked up as TagStandard::RelayMetadata has started adding
a traling slash.
refactor cli output test function `expect_send_with_progress` so that
relays can succeed / fail in a random order
|
|
bump nostr and nostr-sdk packages and also in test_utils
remove custom ncryptsec implementation and use the newly
added implementation nip49 version in rust-nostr
note a patched v0.30 is used so that log_n is exposed so that
user can be warned it might take a few seconds to decrypt.
this has now been merged into the library.
note that this will no longer decrypt existing ncryptsec values as
it is uses a longer string. this should therefore be bundled with
the upcoming change to storing nsec and ncryptsec in git config.
|
|
to repo announcements, patches and cover letters
|
|
fixes:
nostr:nevent1qqsz2g7gexkmqgr0x4g5kf9py8vx06p8nyn78v7mhevprasnyfy5swcpp4mhxue69uhkummn9ekx7mqzyr7jprhgeregx7q2j4fgjmjgy0xfm34l63pqvwyf2acsd9q0mynuzdyll0c
|
|
in addition to being used to create a new proposal revision,
in-reply-to can now be used to reference other events and npubs.
for example an issues or kind 1 threads where the proposal is relevant
the proposal will only be marked as a revision if the first parameter is
a reference to an existing proposal root
|
|
as over the last 2-3 weeks it has consitantly be timing out
|
|
after a new root proposal has been sent
link directly to gitworkshop.dev as well as njump.me until
gitworkshop is added to njump
|
|
seperate connection timeout from get_events_of timeout and reduce it to 3s
|
|
in line with initial merged nip34 spec
|
|
if relays do not connect and retun events within 3s
they will rarely connect at all
this could be reversed when get_events is used async
|
|
rust-nostr author suggested this might be a false positive
https://github.com/rust-nostr/nostr/pull/375#issuecomment-2022245832
|
|
upgrade from rust-nostr v0.27 ~> v0.28 introduces this panic
presumably get_events attempted to write to the MultiProgress
just after it has been removed
|
|
to suppress the clippy too_many_lines in function warning
|
|
to reflect changes in rust-nostr
|
|
to reflect new name in rust-nostr
|
|
to reflect new name in rust-nostr
|