Heh actually that's not an issue, the directory_get_entries documentation does not guarantee an order.
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Feed All Stories
All Stories
All Stories
Sep 22 2021
Sep 22 2021
Possible solution: store them as an ascii string instead of an integer.
Possible solution: store a rank along with each directory entry, but ignore it unless we are reconstructing a git object or computing a SWHID (v1?)
Following what i said in the loader diff, i'm actually closing this.
Ack on the lock folder but i won't attend to it immediately.[1] D6318
swh-public-ci added a comment to D6318: opam: Allow shared state between loader runs using multi-instance opam root.
Build is green
ardumont updated the diff for D6318: opam: Allow shared state between loader runs using multi-instance opam root.
Refactor a bit (drop --strict as suggested)
In T1805#70880, @douardda wrote:
- pagination comes from proper usage of links https://swagger.io/docs/specification/links/
- batches comes from proper usage of parameter serialization https://swagger.io/docs/specification/serialization/
it's true these do not come "for free" but I still have the impression there is an "Open API way" of handling these and we should stick to them.
Harbormaster failed remote builds in B23890: Diff 22960 for D5992: add support for the CVS loader to 'Save Code Now'!
Build has FAILED
Done and dusted.
Upgrade hedgedoc to 1.9.0
olasd committed rSPSITEf5110a49138e: Add apt source configuration for nodejs LTS 14.x (authored by olasd).
Add apt source configuration for nodejs LTS 14.x
Update expired yarn public key
rebased patch
vsellier committed rDDOC3a935af020e7: add a documentation page to list the services urls (authored by vsellier).
add a documentation page to list the services urls
stsp committed rDLDCVSa7eaeb89f46f: eliminate code duplication of logic in process_cvs_changesets() (authored by stsp).
eliminate code duplication of logic in process_cvs_changesets()
vsellier committed rDENV2507d723a1e6: POC a default smaller profile equivalent to the default docker-compose (authored by vsellier).
POC a default smaller profile equivalent to the default docker-compose
vsellier committed rDENV46090bb0b458: Upgrade the registry-ui and fix a CORS issue blocking the access to the registry (authored by vsellier).
Upgrade the registry-ui and fix a CORS issue blocking the access to the registry
swh-public-ci added a comment to D6298: eliminate code duplication of logic in process_cvs_changesets().
Build is green
ardumont committed rDLDBASEf4556e0113fa: opam: Initialize opam root directory outside the constructor (authored by ardumont).
opam: Initialize opam root directory outside the constructor
rebased patch
Following what i said in the loader diff, i'm actually closing this.
Ack on the lock folder but i won't attend to it immediately.
tl; dr Inhibit the cooldown* which default to 7 days.
As it got already scheduled (with whatever results), this is dropped from the grab call.
vault_repro.py: update to latest CLI
add check_consistency.py
LGTM (not checked everything is accurate nor there are obvious missing services, but it's a huge improvement as is, thx)
In T1805#45984, @vlorentz wrote:Items 5, 6, 7 aka pagination, auth and batches - I believe these come naturally with item 4 (specification wise)
They don't. OpenAPI is a specification to describe APIs, and it contains absolutely nothing about pagination or batches.
- Remove useless ending '/'
- define VPN / private meanings
Build is green
I think there was already a problem before, but since we have now more chance to hit it, I'd really like the opam_init process to lock the directory when running opam commands.
In D6165#163629, @vlorentz wrote:What is the reason for this change? Is it more efficient assign requests to workers based on ID rather than randomly?
In D6272#163627, @vlorentz wrote:Why the renaming? And it's an RPC API, not REST
It would be nice to have a README fil in swh/lister/maven/tests/data explaining what the data files are, where they come from, how they have been generated, etc.
In D6273#163625, @vlorentz wrote:Why?
In D6165#163629, @vlorentz wrote:What is the reason for this change? Is it more efficient assign requests to workers based on ID rather than randomly?
ardumont updated the task description for T3584: loader mercurial edge case about missing mapping from revision to hgnode-id.
jayeshv retitled D6291: WIP - Unit tests for the mongo backend from Unit tests for content and directory add methods in mongo to WIP - Unit tests for the mongo backend.
Build is green
- added stubs
- small fixes
- more doc strings
- Better doc strings
ardumont committed rSPSITEcf6550783c19: Drop no longer needed configuration option. (authored by ardumont).
Drop no longer needed configuration option.
ardumont changed the status of T3581: List heptapod instance foss.heptapod.net from Open to Work in Progress.
ardumont moved T3581: List heptapod instance foss.heptapod.net from Backlog to Weekly backlog on the System administration board.
Sep 21 2021
Sep 21 2021
Ok.
What is the motivation for merging directories?
What is the motivation for merging directories?
Sorry, my question was about opam itself. Are you sure it won't corrupt data, or wait on a global lock when accessing that directory?
Sorry, my question was about opam itself. Are you sure it won't corrupt data, or wait on a global lock when accessing that directory?
ardumont renamed T3590: opam loader: Ensure required opam state is shared amongst ingestion/listing runs from opam loader: Ensure the current opam state is shared amongst ingestion to opam loader: Ensure required opam state is shared amongst ingestion/listing runs.
Same comment as on the other diff: looks like it is for consistency with the new rabbitmq RPC, but please explain it
Looks like it is supposed to be replaced by the rabbitmq-based RPC? Please explain it in the commit/diff message.
Can two listers run in parallel while using the same directory?
What is the reason for this change? Is it more efficient assign requests to workers based on ID rather than randomly?
Why the renaming? And it's an RPC API, not REST
Why?
Can two listers run in parallel while using the same directory?
Harbormaster failed remote builds in B23878: Diff 22947 for D6289: vault: Escape provided email address in the URL!
Build has FAILED
Build is green
vlorentz committed rDSTO61e9e4a3232f: cassandra: Make _content_get_from_hashes run concurrently (authored by vlorentz).
cassandra: Make _content_get_from_hashes run concurrently
rebase
fix test + use email-validator + rebase
jayeshv committed rDPROV4c087ea0ece7: Docstring changes for test only methods (Minor). (authored by jayeshv).
Docstring changes for test only methods (Minor).
Build is green
Build is green
Build is green
swh-public-ci added a comment to D6293: load_status should not be 'uneventful' if the snapshot changed.
Build is green