IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
This was after concerns were raised about potential time-of-check to
time-of-use (TOCTOU) issues. The check isn't
actually necessary as `OpenOptions::create` set to `true` will create
only if the file doesn't already exist.
428: Update to go-ipfs 0.7.0 (used in interop tests) r=koivunej a=niklaslong
We were using `0.6.0`, this upgrades to `0.7.0`.
I also removed the `--bits` option as `0.7` uses `ed25519` by default.
Co-authored-by: Niklas Long <niklas@equilibrium.co>
427: Further reduction of ipfs-http compilation times, take two r=aphelionz a=koivunej
This PR supercedes #422 as bors couldn't handle the "PR is from removed repository" case, so this is made from a branch which has only the same commit as the previous PR. PR uses the balanced tree approach from warp issue for further compilation time speedups. The commit hash differs because I apparently applied it per githubs instructions on top of different commit than the original seems to be, with different metadata.
Closes#422.
Co-authored-by: Artem Tarasov <artem@devopps.de>
420: docs(http): initial draft of tutorial r=koivunej a=niklaslong
This is a first draft of the tutorial mentioned in #402.
It currently covers:
- installing rust
- configuration of the `IPFS_PATH`
- `-- init`
- `-- daemon`
- `ipfs id` from the go-ipfs CLI
Co-authored-by: Niklas Long <niklas@equilibrium.co>
Copy-and-paste from https://github.com/seanmonstar/warp/issues/619;
Further reduces compilation time (30-40% improvement observed).
Removed "combine_unify" macro since it didn't add any extra benefit.
423: http: configuration refactor r=niklaslong a=niklaslong
This is a wip but opening as a draft to keep track of this follow up to #421. This is also part of #402.
This introduces the `Config` struct to serve as a facade to the configuration module. It's essentially a flattened `CompatibleConfigFile` struct exposing only the fields necessary to running the daemon.
Co-authored-by: Niklas Long <niklas@equilibrium.co>
421: feat(http): create Profile abstraction to allow port choice r=koivunej a=niklaslong
This PR introduces a `Profile` abstraction to allow the customisation of port selection as discussed in #402.
Two profiles are supported:
- `Test` for use with conformance tests (ephemeral port selection)
- `Default` serves on `4004`
Co-authored-by: Niklas Long <niklas@equilibrium.co>
419: fix(http): change default ipfs path to .rust-ipfs r=koivunej a=niklaslong
This PR changes the default `IPFS_PATH` from `.ipfs` to `.rust-ipfs` as mentioned in #402.
I _think_ this is all that's needed?
Co-authored-by: Niklas Long <niklas@equilibrium.co>
415: add unixfs benchmark r=niklaslong a=niklaslong
Opening this as a draft to have a discussion around the performance of the unixfs adder.
Things I've observed so far:
- benchmark is a little noisy (on my machine at least)
- the adder is significantly slower and more unstable when chunking by small amounts, though this is expected
- I'm not seeing clear spikes for factors of 174, just a general trend towards lower throughput as the input gets larger
- black boxing the returned blocks didn't seem to have much of an impact (small chunking size)
I'm not sure what the best next step would be; as this is meant as a base for future improvements to unixfs, what values/resolution do we want for this?
Co-authored-by: Niklas Long <niklas@equilibrium.co>
408: Fix gha deprecations r=koivunej a=koivunej
In this PR:
- Fixes#407
- "no artifacts found" nagging for missing core files
- 1.47 clippy warning
Github had [a security issue and will be deprecating some functionality](https://github.blog/changelog/2020-10-01-github-actions-deprecating-set-env-and-add-path-commands/). I learned this from the easy-to-miss workflow warnings.
Co-authored-by: Joonas Koivunen <joonas@equilibrium.co>
the warning is quite unuseful since we haven't had a single core dump
after adding it, however there are important warnings from workflows we
should pay attention to.
also, adding a platform limitation cant be done because the ${{ always()
}} seems to have the highest priority.
414: Prepare 0.2.1 release r=koivunej a=koivunej
Tracked in #413. This turned out a bit misc PR with the comment removal from fetch_and_cat and all that.
Co-authored-by: Joonas Koivunen <joonas@equilibrium.co>
