What's Changed

• fix symlink in inputs by @Equanox in #367

Full Changelog: 0.8.2-nix-shell-2...0.8.2-nix-shell-3

This release add an shell directive to the bob.yaml. In case a of an existing shell directive it takes precedence over any given dependencies.

This enables to use pkgs.mkShell inside a nix file... like:

{ pkgs ? import <nixpkgs> {} }:

pkgs.mkShell {
  buildInputs = [
   pkgs.go_1_20
];
}

Example bob.yaml

build:
  build:
    cmd: go build .

shell: shell.nix

This release add an shell directive to the bob.yaml. In case a of an existing shell directive it takes precedence over any given dependencies.

This enables to use pkgs.mkShell inside a nix file... like:

{ pkgs ? import <nixpkgs> {} }:

pkgs.mkShell {
  buildInputs = [
   pkgs.go_1_20
];
}

Example bob.yaml

build:
  build:
    cmd: go build .

shell: shell.nix

Fixes a bug which left a target in an invalid state after restoring from cache.

What's Changed

• Fix build cache restore by @rdnt in #329

Full Changelog: 0.8.1...0.8.2

This bugfix release allows to execute bob without the docker daemon being required (if no docker image is set as target).
It also fixes a bug which created an unexpected cache miss.

What's Changed

• Fix child bobfile target ignoring by @rdnt in #326
• Lazily initialize docker registry client by @rdnt in #328

Full Changelog: 0.8.0...0.8.1

This release introduces "file diffing" for targets. It keeps a directory like node_modules intact on a "cache hit". Therefore running hot-reloading instances for react, vuejs or svelte will not break.

What's Changed

• bob clean: add all cmd by @Equanox in #316
• bob inspect: add buildinfo diff by @Equanox in #317
• Bump github.com/stretchr/testify from 1.8.0 to 1.8.2 by @dependabot in #312
• Hotreloading by @zuzuleinen in #294

Full Changelog: 0.7.2...0.8.0

Changelog

• c170e2f Merge branch 'main' into hotrealoading
• 9ffcf03 Merge branch 'main' into hotrealoading
• 01901b4 Merge branch 'main' into hotrealoading
• 57320fe add cmd to list files
• 3fbc230 add file hash to artifact metadata
• d9d9ffc add inspect buildinfo cmd; add content file hash to buildinfo file
• de0f9c6 assure input hash comparison is similar to input hash during build
• 464b8c0 check file by file when extracting the artifact or invalidating targets
• 13908b5 cleanup
• 3791569 debug
• c229b0d debug buildinfoFiles
• 6749119 dont check filehash on verifyShallow
• a92a4b5 extra debug
• 38ca8c8 fix bob clean targets cmd
• 000502f fix last failing test
• 8b342e4 fix lint
• efda6ff fix path for invalid file cleanup
• 686e946 fix symlink when it has changed file pointer after build
• 0fc028f fix typo
• b388102 if child task has changed check the target of the task
• 00c64f2 ignore node_modules/.cache
• a9abaf8 improve verify shallow performance
• be8ddb2 no need for pointers
• 667167c prevent missuse of cleanup
• 20c5319 readd usage of takID and secure deleetion
• c5d03ad remove bob clean
• 7a421fe remove debug; fix build issue on missing buildinfo
• 62feee5 remove file hash content on artifact meta
• a419571 remove target
• 6c42abd remove typo
• 83e8de9 test bob clean
• 6050334 verify of hash of file contents upon target verification

Changelog

• dd6a2bd Add Bazel comparison
• 6919029 Bump github.com/stretchr/testify from 1.8.0 to 1.8.2 (#312)
• 80f1b1e bob clean: add all cmd
• 8ec09da bob inspect: add buildinfo diff (#317)

Changelog

• 9c872c5 Merge branch 'main' into performance-play
• 5676e8d Perfomance verify duplicate targets (#286)
• 00d98c4 TODO create cache obj for env
• 7957a21 adapt tests
• aad0d12 add benchmark
• b60e6a3 add debug flag to print more build related metrics
• 11f1b22 add more benchmarks
• 0f7992c adjust input tests to be relative
• 702b310 adjust to golang 19
• 93921eb append unique
• bfb891e avoid appen unique
• ce64f16 avoid docker regsitry client generation for each task
• 86356b6 avoid toLower
• c8c41e6 bugfix: prevent send on close channel (#282)
• 871ef14 cleanup
• bfd31de cleanup
• 70c1678 cleanup and fix execution time
• e94249b debug zfs docker
• 7e19a52 debug zfs docker imageSaveToPath
• 21dd861 debug zfs docker remove all except imageSaveToPath
• f1f037d detetct the docker registry storage driver
• 697a4ac distinc clients
• 8c1d161 exchange unique with appendUnique
• 67b445f filter files in parallel
• 1de11ba fine grained metrics
• 3110ecd fix envutil tests
• 69f2fd9 fix input + target tests
• 3ec70a9 fix lint
• f7f5a63 fix lint
• 00c0d5f fix mempry leak & input tests
• 2b83060 ignore invalid symlinks already in ListRecursive
• df649ba no docker logs but filesystem
• 3f32176 nonblocking worker queue
• 407b583 optimize glopping
• e147119 optimized tasks
• dee75b3 parallel filter inputs [tests failing]
• 0eb3a29 performance: nix dependencies (#287)
• e0f8eed pin latest protoc version
• 6180d0e readd summary
• 393b094 refactor nix-builder (#298)
• 52d41e2 remove debug summary
• 5542979 remove filterInput from worker
• ba5341c remove playbook lock
• 1c40c13 revert docker moby mutexes
• acd12ff speedup compute input hash (#289)
• 3ca908e undo
• c002ce5 update flake
• c8d79ff use a in memory env cache
• 49f9683 use docker lock less exaustive
• 2e62046 use optimized tasks
• b2f1725 use unique instead of appendUnique
• 49b795b use warning for symbolic link

This brings performance improvements based on internal optimization.

• Internally the build-graph was walked extensively due to parallel execution this was reduced significantly.
• The env created by nix is reused which reduces the load on garbage collection on big repos.

What's Changed

• bugfix: prevent send on close channel by @Equanox in #282
• perfomance verify duplicate targets by @zuzuleinen in #286
• performance: nix dependencies by @zuzuleinen in #287
• test improve compute input hash by @zuzuleinen in #289
• refactor nix-builder by @Equanox in #298
• performance play by @Equanox in #292

Full Changelog: 0.7.1...0.7.2