GSoC port-util project

Michael Dickens michaelld at macports.org
Mon Feb 15 18:59:19 PST 2016


On Mon, Feb 15, 2016, at 09:33 PM, Mihai Moldovan wrote:
> On 16.02.2016 01:34 AM, Michael Dickens wrote:
> > An example would be a "check and try update" task that would: do a
> > livecheck and if found to be old then automatically try to do the
> > update: change the version, download the distfile, update the checksums
> > in the Portfile, try to upgrade using -s -k to specify from source and
> > to keep stuff around if successful.
> 
> Just a thought on this particular item... maybe it's just me OCDing, but
> I don't
> think a procedure like this should be sanctioned. Blindly updating like
> this is
> dangerous, because you run into problems with also updated dependencies,
> which
> you might not even see.
> 
> I generally diff the old and new version (if possible using SCM as that's
> more
> comfortable) and look through what might have changed within the
> autotools files
> (or whatever build system the port is using.)
> 
> It's a tedious task, but you risk missing changes otherwise.

I somehow doubt that most of us Portfile-level developers do a lot of
SCM diffing between versions. I know for myself that there's just not
enough time in the day to do so for every port I maintain. For some
ports I do do this, just just a few. That said, if any specific MP dev
want to do SCM diffing then that's his/her right IMHO. I don't think
this MO should be required of all MP devs on all ports.

I see this particular item as a quick way to test to see if a port needs
updating, and if so then test to see if it is easily updatable. It has
everything to do with buildability, and nothing to do with diffing
between versions.

All of this said, I think we're talking hypotheticals here anyway. I'm
not even sure this is what pixilla meant in the first place. - MLD


More information about the macports-dev mailing list