'time herdstat --metadata apache' (fresh install, three runs) app-portage/herdstat-1.1.91 dev-cpp/libherdstat-0.2.0 dev-cpp/libebt-1.3.0 real 0m28.667s user 0m0.592s sys 0m0.540s real 0m9.071s user 0m0.416s sys 0m0.348s real 0m10.390s user 0m0.376s sys 0m0.420s vs: app-portage/herdstat-1.1.2 dev-cpp/libherdstat-0.1.1-r1 real 0m0.184s user 0m0.036s sys 0m0.012s real 0m0.045s user 0m0.036s sys 0m0.008s real 0m0.043s user 0m0.032s sys 0m0.012s I've asked another user on #gentoo-bugs to test the latest version and that's what he gets with 1.1.91 on the first run: real 1m5.914s user 0m4.673s sys 0m5.864s Yuck... :(
I hit this too. I've had to mask 1.1.91*, as I use herdstat in all kinds of scripts, and it makes it just unusable.
@compnerd: What is going on with herdstat? Is this a work in progress? Please add some status infos or write me a mail.
There is basically no upstream for this package anymore, unless someone volunteers. I'd recommend for tree-cleaners to look into wiping this out of the tree. There are several open bugs: http://bugs.gentoo.org/buglist.cgi?quicksearch=herdstat
Helps if I CC them...
Uh, punting this is not a viable solution, developers are using this thing frequently, including myself. This 1.1.91 rewrite is screwed, 1.1.2 version is perfectly usable after fixing a couple of minor issues either manually or in the ebuild.
I agree... I use this extensively in my scripts, and would be very sorry to see it go.
Well, something needs to happen with this package. Either punt the new stuff, or fix the old stuff. There are a bunch of bugs open for this for quite some time now. compnerd: opinions?
Is anyone strongly against punting 1.1.91 out of the tree? Its known broken, and probably isn't ever going to be fixed.
Nah, just kill it, thanks. :)
yup, kill it
Punted.
We also need to kick dev-cpp/libherdstat-0.2.0. herdstat-1.1.2 won't build against it and it's causing an up/downgrade cycle.
Removed