It seems there are two opened bugs against this version, hence, maybe this will need to wait... but opening for remembering to stabilize a version that stops using the dead "avahi" USE flag naming in favor of "zeroconf" when we are able to Thanks
This version of root can't be sanely built on systems with >=gcc-5, because bundled llvm have no support for gcc abi tags. Is this ok for stabilization? ATM >=gcc-5 is not yet stable, but root will become a major blocker for that stabilization, because there is no other way to fix this version. The only solution is to update to 6.08.x, which is on my list, but a huge PITA on its own... Please note that root-5 branch is not affected by this problem, so it can be left as stable. I plan to add slotting support for root-5/6, since huge community of scientists still need root-5 and will need it for more years to come. I think it will be better to stabilize root-5.34.36 for now.
(In reply to Pacho Ramos from comment #0) > but opening for remembering to stabilize a version that > stops using the dead "avahi" USE flag naming in favor of "zeroconf" when we > are able to It is already fixed for root-6.04.06. Will fix shortly for older versions. But how on the earth I'm supposed to know that USE=avahi is deprecated in favour of USE=zeroconf? - there is no deprecation comment in profiles/use.desc; - no news item; - no discussion on gentoo-dev ML (at least no messages with "avahi" or "zeroconf" in their topics. Excuse me, but I'm not a telepath :) And please, next time report separate issues as separate bugs.
(In reply to Andrew Savchenko from comment #1) > This version of root can't be sanely built on systems with >=gcc-5, because > bundled llvm have no support for gcc abi tags. Is this ok for stabilization? > ATM >=gcc-5 is not yet stable, but root will become a major blocker for that > stabilization, because there is no other way to fix this version. The only > solution is to update to 6.08.x, which is on my list, but a huge PITA on its > own... Well, gcc >= 5 stabilization is really far as most of the tree is still not compatible, anyway, other option is to rename the USE flag in older ebuilds. That should be easy as it seems no reverse dep would need to get the dep renamed too :D. Maybe this could be done for root-5.34.36 and that version be stabilized. [...] > I think it will be better to stabilize root-5.34.36 for now. Sure! (In reply to Andrew Savchenko from comment #2) [...] > But how on the earth I'm supposed to know that USE=avahi is deprecated in > favour of USE=zeroconf? > > - there is no deprecation comment in profiles/use.desc; > - no news item; > - no discussion on gentoo-dev ML (at least no messages with "avahi" or > "zeroconf" in their topics. > > Excuse me, but I'm not a telepath :) > > And please, next time report separate issues as separate bugs. Because it was discussed many years ago but... as usual, even if approved, nobody went ahead fixing the things. Some months ago I remembered the issue again when needing to enable both variants to get the support and I started to simply go ahead and fix them :/ I think I collected most of the ML items in: https://bugs.gentoo.org/show_bug.cgi?id=477620 But even in my case I don't remember if I included all the links and info there... as it's from 2013 :(
Arch teams, please stabilize =sci-physics/root-5.34.36.
(In reply to Pacho Ramos from comment #3) [...] > Well, gcc >= 5 stabilization is really far as most of the tree is still not > compatible, anyway, other option is to rename the USE flag in older ebuilds. > That should be easy as it seems no reverse dep would need to get the dep > renamed too :D. Maybe this could be done for root-5.34.36 and that version > be stabilized. Done. [...] > Because it was discussed many years ago but... as usual, even if approved, > nobody went ahead fixing the things. Some months ago I remembered the issue > again when needing to enable both variants to get the support and I started > to simply go ahead and fix them :/ > > I think I collected most of the ML items in: > https://bugs.gentoo.org/show_bug.cgi?id=477620 > > But even in my case I don't remember if I included all the links and info > there... as it's from 2013 :( eix tells me there are only 18 packages with USE="avahi" left in the tree. I suggest you to make bug 477620 a tracker and file bugs for each of these packages.
Yeah... that is the current situation after fixing them... even the situation is now better as for most of them we are only waiting for stabilizations... that is the reason I filled this one ;) Thanks for the help!
An automated check of this bug failed - repoman reported dependency errors (23 lines truncated): > dependency.bad sci-physics/root/root-5.34.36.ebuild: PDEPEND: amd64(default/linux/amd64/13.0) ['~app-doc/root-docs-5.34.36[http=,math=]'] > dependency.bad sci-physics/root/root-5.34.36.ebuild: PDEPEND: amd64(default/linux/amd64/13.0/desktop) ['~app-doc/root-docs-5.34.36[http=,math=]'] > dependency.bad sci-physics/root/root-5.34.36.ebuild: PDEPEND: amd64(default/linux/amd64/13.0/desktop/gnome) ['~app-doc/root-docs-5.34.36[http=,math=]']
An automated check of this bug succeeded - the previous repoman errors are now resolved.
For run-time testing please build with these flags enabled: USE="X examples graphviz math opengl tiff" Afterwards run on X-capable system: $ root and type .demo On left panel there will be a list of demos. They should work fine: either open window with some results and/or print data to console. P.S. Should instructions for run-time testing be provided somewhere within ebuild?
amd64 stable
x86 stable. Closing.
