Summary: | sys-app/portage-2.2 --autounmask behavior (automatically edit config) | ||
---|---|---|---|
Product: | Portage Development | Reporter: | Domen Kožar <domen> |
Component: | Enhancement/Feature Requests | Assignee: | Portage team <dev-portage> |
Status: | RESOLVED FIXED | ||
Severity: | enhancement | CC: | esigra, jlec, pacho, tommy |
Priority: | High | Keywords: | InVCS |
Version: | 2.2 | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | 280097 | ||
Bug Blocks: | 300071, 358927 |
Description
Domen Kožar
2010-11-16 16:47:57 UTC
It's worth noting that bug 258371 would produce a similar effect (though only for USE dependencies). (In reply to comment #0) > My suggestion to make it automatic is: > > 1. if --pretend is not passed, populate package.* family automatically and > start emerging the package > 2. if --pretend is passed, ask for confirmation before doing 1. The "ask for confirmation" seems to be the behaviour intended for --ask, not --pretend. (In reply to comment #2) > (In reply to comment #0) > > 1. if --pretend is not passed, populate package.* family automatically and > > start emerging the package > > 2. if --pretend is passed, ask for confirmation before doing 1. > The "ask for confirmation" seems to be the behaviour intended for --ask, not > --pretend. I concur; 1. emerge --autounmask foo => package.* populated, emerge proceeds. 2. emerge -p --autounmask foo => print the changes that would be made to package.*, print the packages that would be emerged 3. emerge -a --autounmask foo => As 2, with "Are you sure?", then as 1. This is in git now: http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=c492b1b3ed631b6802ef1192f59d2ef93967fb0a Why reinvent the wheel with --autounmask-write, if we have standardized --ask parameter? It feels so much intuitive to use global flag to handle logical situations like ask the user for confirmation or just do it. Using only --autounmask to output what needs to be process further and can be totally automatized feels rather redundant to me. Also, --verbose can output the changes needed to be modified in /etc/portage/ if the user want's to know about that. I'm totally thinking from UX perspective and want to double check decision about that. Other than that, this change really made my rainy day. Portage is becoming an masterpiece of package management. (In reply to comment #5) > Why reinvent the wheel with --autounmask-write, if we have standardized --ask > parameter? It's more flexible to use separate options whenever the things that they control are separable. However, I think we should modify the --autounmask-write behavior to respect the --ask option and prompt the user before making the changes. Also, note that you can set EMERGE_DEFAULT_OPTS="--autounmask-write" if you want it enabled by default. Support for --ask is in git now: http://git.overlays.gentoo.org/gitweb/?p=proj/portage.git;a=commit;h=1c16fe8e3740e6b9e1e1562731a0c4b5a000fd92 Thanks! I'll try the changes today and confirm the feature :) This is released in 2.2.0_alpha34. However, I'll leave this bug open until it's released in an umasked version. I plan to release portage-2.1.10 with this feature in approximately 3 weeks. This is fixed in 2.1.10 |