Gentoo Package Repository now using Git

ppurka · Advocate Joined: 26 Dec 2004 Posts: 3256

Update Got my answer. Renamed the gentoo.conf.rsync to .gentoo.conf.rsync and it worked. It seems the rsync configuration was overriding the git configuration. Strange, because I thought it only reads files ending in .conf

I can not seem to sync using git. Can someone explain the reason?

hasufell · Retired Dev Joined: 29 Oct 2011 Posts: 429

I don't know if it can properly initialize the repo if the "/usr/portage" directory already exists. In addition, if it's gone, the "profiles/" subdirectory of the portage tree is missing which means you don't even have a working profile. That can break/confuse some tools.

I'd do the initial clone manually.

edit: tested in a docker environment, it should work if you remove /usr/portage first

ppurka · Advocate Joined: 26 Dec 2004 Posts: 3256

WWWW · Tux's lil' helper Joined: 30 Nov 2014 Posts: 143

I was reading about wtf was $harepoint, and apparently can be used for code versioning.

Maitreya · Guru Joined: 11 Jan 2006 Posts: 441

Yamakuzure · Posted: Wed Sep 16, 2015 8:30 am Post subject:

Just a question:

Is it normal that when ever I sync I get full profile updates?

dweezil-n0xad · Posted: Thu Sep 17, 2015 7:02 am Post subject:

charles17 · Advocate Joined: 02 Mar 2008 Posts: 3664

hasufell · Retired Dev Joined: 29 Oct 2011 Posts: 429

Yamakuzure · Posted: Thu Sep 17, 2015 12:09 pm Post subject:

Oh! I am still using hasufell's postsync stuff, so that might have caused the constant repeating of the update checks, right?
_________________
Important German:

"Aha" - German reaction to pretend that you are really interested while giving no f*ck.
"Tja" - German reaction to the apocalypse, nuclear war, an alien invasion or no bread in the house.

hasufell · Retired Dev Joined: 29 Oct 2011 Posts: 429

steveL · Posted: Sat Sep 19, 2015 7:10 am Post subject:

hasufell · Retired Dev Joined: 29 Oct 2011 Posts: 429

ct85711 · Veteran Joined: 27 Sep 2005 Posts: 1791

miket · Posted: Sat Sep 19, 2015 3:51 pm Post subject:

I wish I had had enough time to review the patching-in technique you described for the older history. Here is the big point, though, one I wish you would have kept in mind: older history matters.

It is not acceptable to have the older history available. Unless that git-replace technique you described can do a seamless merge of the old history, you're not done.

ct85711 · Veteran Joined: 27 Sep 2005 Posts: 1791

miket, think it more of falls down to time vs value. Right now, how I am seeing it, doing all that work doesn't payoff the time spent. Now if this was a development tree, I can see where it would pay off for the extra amount of time however for this tree it doesn't.

Then it also runs into, how many people really need that long of history for the portage tree? Making the assumption that a change now makes something wrong does not also make the past wrong too (especially at the scale you are wanting).

steveL · Posted: Sat Sep 19, 2015 5:17 pm Post subject:

miket · Posted: Sat Sep 19, 2015 5:29 pm Post subject:

I'll concede that not often will people need to look back 10 years, but it does occur quite often that people need that history from just a few months ago. I see that kind of thing a lot in the forums when someone is trying to go back to some version no longer in the tree so that they can unbreak their machines. Or, as in my case, try to figure out where something went wrong so that I can file a meaningful bug report.

Didn't, as I understand, someone already go through the big effort to make that happen? What happened to that?

ct85711 · Veteran Joined: 27 Sep 2005 Posts: 1791

I do know they did try to go through the effort to port the history over, when this was originally thought up. I couldn't really say on what happen on that part, the most I can do is guess on the reasons. I suspect they more of had in issue where they can't get the commits from CVS to line up to git (most likely the case) or more of space usage.

miket · Posted: Sat Sep 19, 2015 9:38 pm Post subject:

desultory · Bodhisattva Joined: 04 Nov 2005 Posts: 9410

alinefr · Posted: Thu Feb 18, 2016 1:29 am Post subject:

Someone in this thread suggested to set:
CACHE_METHOD="/usr/portage/ parse|ebuild*"

... as eix-update was not working by default when 'sync-type = git'. But it was taking ages to build the database. So I did try different values for CACHE_METHOD, and in the end 'assign' is working as it should. A lot more faster than parse|ebuild:

So my CACHE_METHOD is set like this now:
CACHE_METHOD="/usr/portage/ assign"

ExecutorElassus · Posted: Tue Oct 11, 2016 9:52 pm Post subject:

I have /usr/portage/distfiles as a separate partition. When I try to sync, I get an error:

grumblebear · Apprentice Joined: 26 Feb 2008 Posts: 202

Your problem is the initial cloning? Just do:

umount /usr/portage/distfiles
emerge --sync
mount /usr/portage/distfiles

ExecutorElassus · Posted: Wed Oct 12, 2016 5:30 pm Post subject:

It was just the initial cloning, but then a dev told me to move what used to be /usr/portage to /var/portage/repos and distfiles to /var/portage/distfiles, so that it's no longer within the tree that git syncs to. I had to move mountpoints around and edit my fstab, but now it works.

But thanks for the tip. It's good to know I can also just unmount the partition.

Cheers,

EE