[pacman-dev] hackish --ask option ?
Another interesting post from the forum : http://bbs.archlinux.org/viewtopic.php?id=35901 The documentation of the --ask option is quite funny indeed : --ask number Pre-specify answers to questions. It is doubtful whether this option even works, so I would not recommend using it. TODO: docu- ment this more, as I have no idea how it works or when you would use it, or if we should just dump it. I made an attempt of understanding what this option did, and then documenting it, but I doubt it's clear enough. I'm not sure what should be done here : 1) trying to document this option better 2) trying to implement this feature in a better way 3) just dropping it (not sure that's a good idea, since it's apparently useful in some cases). I think that's also low priority (ie not needed for 3.1), but feel free to comment on this issue.
On 8/16/07, Xavier <shiningxc@gmail.com> wrote:
Another interesting post from the forum : http://bbs.archlinux.org/viewtopic.php?id=35901
The documentation of the --ask option is quite funny indeed : --ask number Pre-specify answers to questions. It is doubtful whether this option even works, so I would not recommend using it. TODO: docu- ment this more, as I have no idea how it works or when you would use it, or if we should just dump it.
I made an attempt of understanding what this option did, and then documenting it, but I doubt it's clear enough.
I'm not sure what should be done here : 1) trying to document this option better 2) trying to implement this feature in a better way 3) just dropping it (not sure that's a good idea, since it's apparently useful in some cases).
I think that's also low priority (ie not needed for 3.1), but feel free to comment on this issue.
Anyone against dropping it for purposes of being KISS? I think you are really relying on consistent pacman behavior which I can assure you does not always happen. Or at least drop the current implementation and let someone fix it up to something that makes sense as proposal #2 states. -Dan
Hello, Na Sat, Aug 18, 2007 at 03:04:30PM -0400, Dan McGee <dpmcgee@gmail.com> pisal(a):
Anyone against dropping it for purposes of being KISS? I think you are really relying on consistent pacman behavior which I can assure you does not always happen.
Or at least drop the current implementation and let someone fix it up to something that makes sense as proposal #2 states.
see the "Transaction Conversations" section of alpm.h, why is it hackish? or if we think that manually ORing the options is hackish, then why not removing the --debug option, too? - VMiklos
On Thu, Aug 23, 2007 at 04:51:53PM +0200, VMiklos wrote:
Hello,
Na Sat, Aug 18, 2007 at 03:04:30PM -0400, Dan McGee <dpmcgee@gmail.com> pisal(a):
Anyone against dropping it for purposes of being KISS? I think you are really relying on consistent pacman behavior which I can assure you does not always happen.
Or at least drop the current implementation and let someone fix it up to something that makes sense as proposal #2 states.
see the "Transaction Conversations" section of alpm.h, why is it hackish?
Actually, that's partly what I was asking.
or if we think that manually ORing the options is hackish, then why not removing the --debug option, too?
It was removed as well. A while ago, we actually had a discussion about it, I complained that I found the --debug option was not configurable enough. That only applied to arch pacman, not fw one. I personally don't have problems with ORing options, especially if these options are only used by developers.
Hello, Na Thu, Aug 23, 2007 at 05:04:31PM +0200, Xavier <shiningxc@gmail.com> pisal(a):
or if we think that manually ORing the options is hackish, then why not removing the --debug option, too?
It was removed as well.
ah ok, then sorry for not being up to date :) - VMiklos
participants (3)
-
Dan McGee
-
VMiklos
-
Xavier