svn commit: r410453 - in head/math/fityk: . files

[Alexey Dokuchaev]

On Tue, Mar 08, 2016 at 04:25:30PM +0100, Mathieu Arnold wrote:
> +--On 8 mars 2016 16:06:59 +0100 Pietro Cerutti <gahr at gahr.ch> wrote:
> | So, danfe@, would you mind explaining to me on which basis muting MKDIRs
> | is "at one's discretion" while muting INSTALLs is "actually wrong"?
> | 
> | Given the evidence pointed at by mat@, either both muting MKDIR and
> | INSTALL commands is at one's discretion, or neither is.
> 
> Ok, the IMPORTANT bit here, is that the *install* phase is the most
> important one from the end-user point of view.
> 
> It's where things go live, and his system is modified.

Not in the post-staging era, where *install* phase modifies bits under the
$stagedir.  For package users it's even less important for obvious reasons.

Basically, making those commands vocal is beneficial for us porters and
maintainers, not for the "end-users".

> So we want to show up things thare are modifying his system.  All of them.
> 
> Some argue that mkdir are not that important and can be muted, I kinda
> disagree.

As they say, we agree to disagree. ;-)

> And as the doc is how portmgr thinks things should be (in that case, it's
> been a couple of years where that means what *I* think), well, it's how I
> documented it.

Sorry, but that hardly backs up that what you "kinda disagree" about should
make its way into official documentation so easily.

./danfe