Robot/watchexec
1
0
Fork 0
mirror of https://github.com/watchexec/watchexec.git synced 2024-11-16 17:18:30 +01:00
Code Issues Projects Releases Packages Wiki Activity
watchexec/crates/cli/Cargo.toml

127 lines
3.6 KiB
TOML
Raw Normal View History

Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
[package]
name = "watchexec-cli"
release: cli v1.20.5 (#374) This is a release PR for **cli** to version **1.20.5**. Upon merging, this will automatically build the CLI and create a GitHub release. You still need to manually publish the cargo crate. ``` $ cd crates/cli $ cargo publish ``` To merge this release, review the changes then say: | bors r+ p=10 | |:-:| --- _Edit release notes into the section below:_ <!-- do not change or remove this heading --> ### Release notes - Fix: mimalloc was set as the global allocator in the watchexec library, instead of leaving it to downstreams. (#369) - Releng: s390x and ppc64le targets were added to be prebuilt. (#373)
2022-08-04 01:10:41 +02:00
version = "1.20.5"
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
authors = ["Matt Green <mattgreenrocks@gmail.com>", "Félix Saparelli <felix@passcod.name>"]
license = "Apache-2.0"
description = "Executes commands in response to file modifications"
keywords = ["watcher", "filesystem", "cli", "watchexec"]
categories = ["command-line-utilities"]
Review readmes, add website links
2021-07-21 14:13:59 +02:00
documentation = "https://watchexec.github.io/docs/#watchexec"
homepage = "https://watchexec.github.io"
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
repository = "https://github.com/watchexec/watchexec"
Split readmes
2021-05-08 17:39:13 +02:00
readme = "README.md"
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
Opt in to 2021 ed
2021-10-19 13:51:33 +02:00
edition = "2021"
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
[[bin]]
name = "watchexec"
path = "src/main.rs"
[dependencies]
Upgrade to tokio-console 0.1
2021-12-23 14:28:21 +01:00
console-subscriber = { version = "0.1.0", optional = true }
Ignore homedir if discovered as origin unless explicitly requested Fixes #276
2022-04-12 03:30:39 +02:00
dirs = "4.0.0"
Find and load all ignores for watchexec cli
2021-10-13 13:38:56 +02:00
dunce = "1.0.2"
Update deps
2022-01-18 10:14:34 +01:00
futures = "0.3.17"
Upgrade to miette 4 (#303)
2022-06-11 09:55:43 +02:00
miette = { version = "4.7.1", features = ["fancy"] }
Add experimental --notify option See #139
2021-07-30 19:11:42 +02:00
notify-rust = "4.5.2"
Find and load all ignores for watchexec cli
2021-10-13 13:38:56 +02:00
tracing = "0.1.26"
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
[dependencies.clap]
Upgrade to clap 3 (#304) This is just the upgrade, including the new option categories. Move to the derive API or more advanced features is left for later.
2022-06-11 10:05:11 +02:00
version = "3.1.18"
features = ["cargo", "wrap_help"]
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
Split into more crates (#307)
2022-06-15 05:25:05 +02:00
[dependencies.ignore-files]
release: ignore-files v1.0.0 (#318) Co-authored-by: github-actions <github-actions@github.com>
2022-06-16 13:09:16 +02:00
version = "1.0.0"
Split into more crates (#307)
2022-06-15 05:25:05 +02:00
path = "../ignore-files"
[dependencies.project-origins]
release: project-origins v1.1.0 (#385) This is a release PR for **project-origins** to version **1.1.0**. Upon merging, you will still need to manually publish the cargo crate. ``` $ cd crates/project-origins $ cargo publish ``` To merge this release, review the changes then say: | bors r+ p=10 | |:-:|
2022-08-24 11:23:52 +02:00
version = "1.1.0"
Split into more crates (#307)
2022-06-15 05:25:05 +02:00
path = "../project-origins"
Almost there
2022-06-16 10:17:04 +02:00
[dependencies.watchexec]
release: watchexec v2.0.1 (#390) Co-authored-by: github-actions <github-actions@github.com>
2022-09-07 02:57:50 +02:00
version = "2.0.1"
Almost there
2022-06-16 10:17:04 +02:00
path = "../lib"
[dependencies.watchexec-filterer-globset]
release: filterer-globset v1.0.0 (#335) This is a release PR for **filterer-globset** to version **1.0.0**. Upon merging, you will still need to manually publish the cargo crate. ``` $ cd crates/filterer/globset $ cargo publish ``` To merge this release, review the changes then say: | bors r+ p=10 | |:-:|
2022-06-23 05:33:20 +02:00
version = "1.0.0"
Almost there
2022-06-16 10:17:04 +02:00
path = "../filterer/globset"
[dependencies.watchexec-filterer-tagged]
release: filterer-tagged v0.1.0 (#336) This is a release PR for **filterer-tagged** to version **0.1.0**. Upon merging, you will still need to manually publish the cargo crate. ``` $ cd crates/filterer/tagged $ cargo publish ``` To merge this release, review the changes then say: | bors r+ p=10 | |:-:|
2022-06-23 11:24:07 +02:00
version = "0.1.0"
Almost there
2022-06-16 10:17:04 +02:00
path = "../filterer/tagged"
Reduce featureset of dependencies (tokio,git2)
2021-10-15 03:27:17 +02:00
[dependencies.tokio]
Replace event queue with a priority queue (#302) Solves several issues, generally through delivering signals before filesystem events, preventing situations where an overwhelming amount of events makes it impossible to quit. Does _not_ solve the problem of a queue full of lower-priority events not accepting an urgent message, but that's a rarer issue that's more complicated to overcome. Changes the Filterer trait: adds Priority to `check_event()` Makes some events unfilterable (Urgent priority): SIGINT, SIGTERM, and CTRL_C to the main process. These still need to be handled by `on_action` to do anything, but cannot be stopped before reaching that.
2022-06-11 08:43:11 +02:00
version = "1.19.2"
Reduce deps featureset
2021-10-16 06:25:21 +02:00
features = [
Reduce featureset of dependencies (tokio,git2)
2021-10-15 03:27:17 +02:00
"fs",
Reduce deps featureset
2021-10-16 06:25:21 +02:00
"io-std",
"process",
"rt",
"rt-multi-thread",
"signal",
"sync",
Reduce featureset of dependencies (tokio,git2)
2021-10-15 03:27:17 +02:00
]
Upgrade tracing-subscriber to 0.3
2021-12-23 14:35:26 +01:00
[dependencies.tracing-subscriber]
Update deps
2022-01-18 10:14:34 +01:00
version = "0.3.6"
Upgrade tracing-subscriber to 0.3
2021-12-23 14:35:26 +01:00
features = [
"env-filter",
"fmt",
Add --log-file so logs are easier to collect (#321)
2022-06-16 17:56:59 +02:00
"json",
Fix logging initialisation (#348) - Fixes #345 (first crash, related to `RUST_LOG` being set) - Adds multiple warnings when non-standard logging or build options are set - Fixes the (internal) crate name of the cli so that the log target `watchexec_cli` works - Fixes the verbosity config to target the actual crate names (snake vs kebab case) - Fixes a usage error related to tracing's `log` compatibility
2022-06-28 14:21:19 +02:00
"tracing-log",
Upgrade tracing-subscriber to 0.3
2021-12-23 14:35:26 +01:00
]
Use mimalloc for musl builds
2021-10-15 03:27:32 +02:00
[target.'cfg(target_env = "musl")'.dependencies]
mimalloc = "0.1.26"
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
[build-dependencies]
embed-resource = "1.6.1"
Add tokio-console
2021-09-28 14:47:18 +02:00
[features]
Document feature
2022-02-04 04:46:49 +01:00
## For debugging only: enables the Tokio Console.
Add tokio-console
2021-09-28 14:47:18 +02:00
dev-console = ["console-subscriber"]
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
[package.metadata.binstall]
Update binstall metadata (#383)
2022-08-24 09:19:02 +02:00
pkg-url = "{ repo }/releases/download/cli-v{ version }/watchexec-{ version }-{ target }.{ archive-format }"
bin-dir = "watchexec-{ version }-{ target }/{ bin }{ binary-ext }"
Split into two crates: lib and cli That has a number of advantages: - #193 the build.rs is only run for the CLI, so the Windows manifest is not embedded in the library anymore, opening it up for downstreams. - it sets the stage for decoupling the version numbers of the CLI and library, to have the library increase its major more often, while the CLI retains compatibility further… that is, to have both follow semver - it removes the CLI-only dependencies from the library - it makes compilation a bit faster as compiling the library and the CLI's other dependencies can happen in parallel One major disadvantage: - installing via cargo changes from watchexec to watchexec-cli. Most installs are from prebuilt and from packages, but that's still a potential stumble. And of course, the CLI APIs in the library are gone (they were already deprecated, though). We also take this opportunity to get rid of the clear_screen code and use our new clearscreen library. #99 #171 #185
2021-04-30 17:27:17 +02:00
pkg-fmt = "txz"
[package.metadata.binstall.overrides.x86_64-pc-windows-msvc]
pkg-fmt = "zip"
Use cargo-deb to replace existing deb kludge
2021-06-27 13:55:40 +02:00
[package.metadata.deb]
maintainer = "Félix Saparelli <felix@passcod.name>"
Fix CLI release workflow (#339, #340, #341, #342, #343)
2022-06-23 12:29:04 +02:00
license-file = ["../../LICENSE", "0"]
Use cargo-deb to replace existing deb kludge
2021-06-27 13:55:40 +02:00
section = "utility"
Manually set depends for deb and rpm packages (#358) Should fix #292
2022-06-29 03:28:29 +02:00
depends = "libc6, libgcc-s1" # not needed for musl, but see below
Use cargo-deb to replace existing deb kludge
2021-06-27 13:55:40 +02:00
# conf-files = [] # look me up when config file lands
assets = [
Fix bin filename for deb/rpm packaging (#364) Unsure if deb ever supported that but rpm sure does not. Fixes #292
2022-07-01 15:39:43 +02:00
["../../target/release/watchexec", "usr/bin/watchexec", "755"],
Move config creation out of arg parsing, and start on using libv2
2021-08-24 10:23:37 +02:00
["README.md", "usr/share/doc/watchexec/README", "644"],
Fix CLI release workflow (#339, #340, #341, #342, #343)
2022-06-23 12:29:04 +02:00
["../../doc/watchexec.1.html", "usr/share/doc/watchexec/watchexec.1.html", "644"],
["../../doc/watchexec.1", "usr/share/man/man1/watchexec.1.html", "644"],
["../../completions/zsh", "usr/share/zsh/site-functions/_watchexec", "644"],
["../../doc/logo.svg", "usr/share/icons/hicolor/scalable/apps/watchexec.svg", "644"],
Use cargo-deb to replace existing deb kludge
2021-06-27 13:55:40 +02:00
]
Add rpm support with cargo-generate-rpm
2021-06-27 13:55:56 +02:00
[package.metadata.generate-rpm]
assets = [
Fix bin filename for deb/rpm packaging (#364) Unsure if deb ever supported that but rpm sure does not. Fixes #292
2022-07-01 15:39:43 +02:00
{ source = "../../target/release/watchexec", dest = "/usr/bin/watchexec", mode = "755" },
Fix CLI release workflow (#339, #340, #341, #342, #343)
2022-06-23 12:29:04 +02:00
{ source = "README.md", dest = "/usr/share/doc/watchexec/README", mode = "644", doc = true },
{ source = "../../doc/watchexec.1.html", dest = "/usr/share/doc/watchexec/watchexec.1.html", mode = "644", doc = true },
{ source = "../../doc/watchexec.1", dest = "/usr/share/man/man1/watchexec.1.html", mode = "644" },
{ source = "../../completions/zsh", dest = "/usr/share/zsh/site-functions/_watchexec", mode = "644" },
{ source = "../../doc/logo.svg", dest = "/usr/share/icons/hicolor/scalable/apps/watchexec.svg", mode = "644" },
Move config creation out of arg parsing, and start on using libv2
2021-08-24 10:23:37 +02:00
# set conf = true for config file when that lands
Add rpm support with cargo-generate-rpm
2021-06-27 13:55:56 +02:00
]
Manually set depends for deb and rpm packages (#358) Should fix #292
2022-06-29 03:28:29 +02:00
auto-req = "disabled"
# technically incorrect when using musl, but these are probably
# present on every rpm-using system, so let's worry about it if
# someone asks.
[package.metadata.generate-rpm.requires]
glibc = "*"
libgcc = "*"
Reference in a new issue Copy permalink