cave-resume — Resume a failed resolution from 'cave resolve'
cave resume --resume-file state [ --retry-failed ] [ --retry-skipped ]
Resumes a failed resultion from 'cave resolve'. To enable resumes, use 'cave resolve --resume-file state --execute', and then if errors occur, use 'cave resume --resume-file state [ --retry-failed ] [ --retry-skipped ]' to try again.
Global options, used by every subcommand.
- -h , --help
- display help message
Retry options. By default, 'cave resume' will treat packages that were already skipped or already failed as skipped or failed, respectively.
- -r , --retry-failed (+r , --no-retry-failed)
- Retry any job that has already failed
- -R , --retry-skipped (+R , --no-retry-skipped)
- Retry any job that has already been skipped. Note that the job will just be skipped again unless circumstances have changed.
- -s , --skip-failed (+s , --no-skip-failed)
- Skip any job that has already failed
Failure handling options.
- -C , --continue-on-failure
Whether to continue after an error occurs
- never (n)
- Never (default)
- if-satisfied (s)
- If remaining packages' dependencies are satisfied
- if-independent (i)
- If remaining packages do not depend upon any failing package
- always (a)
- Always (dangerous)
- --resume-file (--no-resume-file)
- Write resume information to the specified file. If a build fails, or if '--execute' is not specified, then 'cave resume' can resume execution from this file.
Options controlling jobs and parallelism.
- -f , --fetch (+f , --no-fetch)
- Skip any jobs that are not fetch jobs. Should be combined with --continue-on-failure if any of the packages to be merged have fetch dependencies.
- -J , --fetch-jobs
- The number of parallel fetch jobs to launch. If set to 0, fetches will be carried out sequentially with other jobs. Values higher than 1 are currently treated as being 1. Defaults to 1, or if --fetch is specified, 0.
Options controlling which phases to execute. No sanity checking is done, allowing you to shoot as many feet off as you desire. Phase names do not have the src_, pkg_ or builtin_ prefix, so 'init', 'preinst', 'unpack', 'merge', 'strip' etc.
- Skip the named phases
- Abort when a named phase is encountered
- Skip every phase until a named phase is encountered
Control to which package or packages these phase options apply
- All packages (default)
- Only the first package on the list
- Everything except the first package on the list
- Only the last package on the list
- Everything except the last package on the list
- Only packages that are targets
- Only packages that are not targets
Options controlling which programs are used to carry out various tasks. Any replacement to the standard program must provide exactly the same interface. In all cases, $CAVE can be used to get the path of the main 'cave' executable. Note that unless an option is explicitly specified, an internal implementation of the default command might be used instead of spawning a new process.
- The program used to display the resolution. Defaults to '$CAVE display-resolution'.
- The program used to graph jobs. Defaults to '$CAVE graph-jobs'.
- The program used to execute the resolution. Defaults to '$CAVE execute-resolution'.
- The program used to perform actions. Defaults to '$CAVE perform'.
- The program used to perform world updates. Defaults to '$CAVE update-world'.
- The program used to create Graphviz graphs. Defaults to 'dot'.