On Mon, Mar 10, 2008 at 6:29 PM, Dan McGee wrote:

...

...

...

From now on 'pacman -Sc target' will first clean the cache, then install target...

This seems sane, I'm just not sure how useful (or expected) this behavior would be. Can you give a use case?

I actually think this could cause confusion. A user could very well think that pacman is going to clean only the target package.

Scott

To be honest, the main reason for this patch was to clean needs_transaction() and needs_root() check. See: http://www.archlinux.org/pipermail/pacman-dev/2008-March/011382.html However, this wasn't just selfish: * the current code also can be fooled easily: -Sc needs root and clean cache -Scg DOESN'T need root and [tries to] clean cache only * We have no clear definition what will happen on pacman -Syucg foo (*)[this is a syntactically correct command-line now], so imho this patch doesn't hurt anything: users still should use 'pacman -Sc' "alone", but power-users can use it in combination with -Su to clean-up cache (to get some disk space) and THEN do a sysupgrade. * answer to your fear: this won't be mentioned in manual and see (*): these command-lines are already allowed with undefined result. Bye ---------------------------------------------------- SZTE Egyetemi Könyvtár - http://www.bibl.u-szeged.hu This mail sent through IMP: http://horde.org/imp/

2008/3/11 Nagy Gabor :

...

On Mon, Mar 10, 2008 at 6:29 PM, Dan McGee wrote:

...

...

...

From now on 'pacman -Sc target' will first clean the cache, then install target...

This seems sane, I'm just not sure how useful (or expected) this behavior would be. Can you give a use case?

I actually think this could cause confusion. A user could very well think that pacman is going to clean only the target package.

Scott

To be honest, the main reason for this patch was to clean needs_transaction() and needs_root() check. See: http://www.archlinux.org/pipermail/pacman-dev/2008-March/011382.html However, this wasn't just selfish: * the current code also can be fooled easily: -Sc needs root and clean cache -Scg DOESN'T need root and [tries to] clean cache only * We have no clear definition what will happen on pacman -Syucg foo (*)[this is a syntactically correct command-line now], so imho this patch doesn't hurt anything: users still should use 'pacman -Sc' "alone", but power-users can use it in combination with -Su to clean-up cache (to get some disk space) and THEN do a sysupgrade. * answer to your fear: this won't be mentioned in manual and see (*): these command-lines are already allowed with undefined result.

Well perhaps we should address the real problem then and find a better way to allow/disallow options in combination with each other? It scares me more to say "oh, the behavior is weird, so we aren't going to even put it in the manpage". -Dan

On Mon, Mar 10, 2008 at 9:28 PM, Scott Horowitz wrote:

I actually think this could cause confusion. A user could very well think that pacman is going to clean only the target package.

I agree with Scott here. If someone were to tell me that "pacman -Sc foobar" worked, I would assume it'd clean up old versions of foobar, not clean the cache and sync foobar.

2008/3/12 Nagy Gabor :

...

On Mon, Mar 10, 2008 at 9:28 PM, Scott Horowitz wrote:

...

I actually think this could cause confusion. A user could very well think that pacman is going to clean only the target package.

I agree with Scott here. If someone were to tell me that "pacman -Sc foobar" worked, I would assume it'd clean up old versions of foobar, not clean the cache and sync foobar.

I just repeat myself: Currently this works, and cleans the whole cache and stops.

Well then I'll repeat myself- we should fix the options parsing. -Dan