The Rust Log Analyzer (RLA) is a tool that analyzes the CI build logs of the rust-lang/rust repository with the purpose of automatically extracting error messages from failed builds.
This repository contains three components:
rust_log_analyzer library, which contains the analysis logic.rla-offline binary, a collection of various tools to run the analyzer off-line (see the output of rla-offline --help for available commands).rla-server binary, a web server which receives GitHub webhooks and automatically posts analysis results to the Rust repository.RLA uses the log crate to write all output. By default, anything logged at INFO or higher will be printed. You can change this behavior by setting the RLA_LOG environment variable, using the syntax specified by the env_logger crate.
To run commands which access online resources, you have to provide the required authentication information in environment variables:
GITHUB_TOKEN to a personal access token with at least "repo" scope.To initialize a new index file, perform the following steps:
rla-offline dl command.
release mode.rla-offline dl --ci actions --repo rust-lang/rust -c 40 --branch auto --passed -o data/trainingrla-offline learn command.
rla-offline learn --ci actions -i demo.idx data/trainingrla-offline dl command or manually from your CI provider.
.brotli, or assume uncompressed data otherwise.rla-offline extract-one command analyze the log file.
rla-offline extract-one --ci actions -i demo.idx my-log.txtNote: This process will / should be integrated as regression tests.
rla-offline dl --ci actions --repo rust-lang/rust -c 50 --failed -o data/failedrla-offline extract-dir command to analyze all the log files and write the results to a separate directory.
rla-offline extract-dir --ci actions -i demo.idx -s data/failed -d data/errThe index file can be stored either in the local filesystem (by providing the
absolute or relative path to the file) or in S3 (by providing a
s3://{bucket}/{key} URL). The S3 region of the bucket is detected
automatically at startup.
Rust
98.2%
Dockerfile
1.8%
The Rust Log Analyzer (RLA) is a tool that analyzes the CI build logs of the rust-lang/rust repository with the purpose of automatically extracting error messages from failed builds.
This repository contains three components:
rust_log_analyzer library, which contains the analysis logic.rla-offline binary, a collection of various tools to run the analyzer off-line (see the output of rla-offline --help for available commands).rla-server binary, a web server which receives GitHub webhooks and automatically posts analysis results to the Rust repository.RLA uses the log crate to write all output. By default, anything logged at INFO or higher will be printed. You can change this behavior by setting the RLA_LOG environment variable, using the syntax specified by the env_logger crate.
To run commands which access online resources, you have to provide the required authentication information in environment variables:
GITHUB_TOKEN to a personal access token with at least "repo" scope.To initialize a new index file, perform the following steps:
rla-offline dl command.
release mode.rla-offline dl --ci actions --repo rust-lang/rust -c 40 --branch auto --passed -o data/trainingrla-offline learn command.
rla-offline learn --ci actions -i demo.idx data/trainingrla-offline dl command or manually from your CI provider.
.brotli, or assume uncompressed data otherwise.rla-offline extract-one command analyze the log file.
rla-offline extract-one --ci actions -i demo.idx my-log.txtNote: This process will / should be integrated as regression tests.
rla-offline dl --ci actions --repo rust-lang/rust -c 50 --failed -o data/failedrla-offline extract-dir command to analyze all the log files and write the results to a separate directory.
rla-offline extract-dir --ci actions -i demo.idx -s data/failed -d data/errThe index file can be stored either in the local filesystem (by providing the
absolute or relative path to the file) or in S3 (by providing a
s3://{bucket}/{key} URL). The S3 region of the bucket is detected
automatically at startup.
Rust
98.2%
Dockerfile
1.8%