Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!

Bug 150251

Summary: sys-devel/binutils: USE=multislot conflicts with new portage-2.1.2_pre2-r4 slotted behavior
Product: Gentoo Linux Reporter: Duncan <1i5t5.duncan>
Component: [OLD] Core systemAssignee: Portage team <dev-portage>
Status: RESOLVED FIXED    
Severity: normal CC: jakub
Priority: High Keywords: InVCS, REGRESSION
Version: 2006.1   
Hardware: All   
OS: Linux   
See Also: https://bugs.gentoo.org/show_bug.cgi?id=150361
Whiteboard:
Package list:
Runtime testing required: ---
Bug Depends on:    
Bug Blocks: 147007    
Attachments: emerge --info
don't pull unavailable slots in with system and world

Description Duncan 2006-10-06 04:43:07 UTC
$emerge -pu world

These are the packages that would be merged, in order:

Calculating world dependencies          u
!!! Ebuilds for the following packages are either all
!!! masked or don't exist:
sys-devel/binutils:x86_64-pc-linux-gnu-2.17 sys-devel/binutils:x86_64-pc-linux-gnu-2.16.1

... done!

$equery l binutils
[ Searching for package 'binutils' in all categories among: ]
 * installed packages
[I--] [ -] sys-devel/binutils-2.16.1-r3 (x86_64-pc-linux-gnu-2.16.1)
[I--] [ ~] sys-devel/binutils-2.17 (x86_64-pc-linux-gnu-2.17)
[I--] [ ~] sys-devel/binutils-config-1.9-r2 (0)

$emerge -pv binutils

These are the packages that would be merged, in order:

Calculating dependencies ... done!
[ebuild   R   ] sys-devel/binutils-2.17  USE="multislot (-multitarget) -nls -test -vanilla" 0 kB

Total size of downloads: 0 kB

$equery l portage
[ Searching for package 'portage' in all categories among: ]
 * installed packages
[I--] [  ] app-portage/portage-manpages-20060913 (0)
[I--] [ ~] sys-apps/portage-2.1.2_pre2-r4 (0)

This is, I believe, due to the new emerge -u slotted behavior,
bugs #4698 and #93469.  As such, emerge --info shouldn't be needed,
given the above command outputs.  However, if it is, just ask.
(I'm on amd64 if it matters.)

Duncan
Comment 1 yogeshbug 2006-10-06 04:45:31 UTC
Pls give emerge --info 
Comment 2 Duncan 2006-10-06 05:20:26 UTC
Created attachment 98929 [details]
emerge --info

emerge --info (minus certain local fs paths info, this /is/ a public site, after all, and that's on a need-to-know basis)
Comment 3 Jakub Moc (RETIRED) gentoo-dev 2006-10-06 10:10:08 UTC
Nice junk in C[XX]FLAGS... Kindly don't report any compile bugs with this. 


See this patch: http://dev.gentoo.org/~zmedico/portage/branches/2.1/patches/installed_deps/installed_deps-2.1.2_pre2-r4.patch
Comment 4 SpanKY gentoo-dev 2006-10-06 10:12:40 UTC
> Nice junk in C[XX]FLAGS... Kindly don't report any compile bugs with this. 

off topic and has nothing to do with the issue at hand
Comment 5 Zac Medico gentoo-dev 2006-10-06 11:26:20 UTC
This is fixed in 2.1.2_pre2-r5.

*** This bug has been marked as a duplicate of 48195 ***
Comment 6 Duncan 2006-10-06 17:30:12 UTC
Thanks, and on the CFLAGS, don't worry, if I can add to CFLAGS, I can just as easily remove from CFLAGS, as routinely do before bug reports or as necessary. =8^) That said, -combine and FEATURES=confcache are the only ones that seem to cause problems, and that's seldom enough I can deal with it when it occurs.  A couple of the others are known to have issues on x86, but I'm not on x86, or I'd likely not be using them.
Comment 7 Zac Medico gentoo-dev 2006-10-06 18:03:56 UTC
Note that the fix for bug 48195 doesn't solve all issues with multislot packages.  For example, some of the slots may not update automatically with world like one would hope and will not be remerged via --emptytree.  What's really needed is some new type of ebuild metadata indicating dynamic slots that may be available.
Comment 8 Zac Medico gentoo-dev 2006-10-06 18:05:33 UTC
I meant to leave this as a duplicate.  Proper dynamic slots support should be proposed via a GLEP.

*** This bug has been marked as a duplicate of 48195 ***
Comment 9 Jakub Moc (RETIRED) gentoo-dev 2006-10-07 00:50:25 UTC
(In reply to comment #6)
> Thanks, and on the CFLAGS, don't worry, if I can add to CFLAGS, I can just as
> easily remove from CFLAGS, as routinely do before bug reports

Not even much funny... You are asking for your compile bugs to be marked INVALID on the fly, probably.

(In reply to comment #7)
> Note that the fix for bug 48195 doesn't solve all issues with multislot
> packages.  For example, some of the slots may not update automatically with
> world like one would hope and will not be remerged via --emptytree.  What's
> really needed is some new type of ebuild metadata indicating dynamic slots 

Seems to work nicely, actually:

# emerge -uDpv world

These are the packages that would be merged, in order:

Calculating world dependencies... done!
[ebuild     U ] sys-devel/binutils-2.17.50.0.3-r1 [2.17.50.0.3] USE="multislot -multitarget nls -test -vanilla" 0 kB 

# qlist -CIev binutils
sys-devel/binutils-2.17.50.0.3
sys-devel/binutils-2.17.50.0.5

 # emerge -epv system | grep binutils
[ebuild  N    ] sys-devel/binutils-config-1.9-r2  0 kB 
[ebuild  N    ] sys-devel/binutils-2.17.50.0.5  USE="multislot -multitarget nls -test -vanilla" 0 kB 
[ebuild  N    ] sys-devel/binutils-2.17.50.0.3-r1  USE="multislot -multitarget nls -test -vanilla" 0 kB 
Comment 10 Zac Medico gentoo-dev 2006-10-08 12:08:33 UTC
We've got the same problem again in 2.1.2_pre2-r6 due to out masking check not being sort of a hack (bug 149816).
Comment 11 Zac Medico gentoo-dev 2006-10-08 14:02:05 UTC
Created attachment 99162 [details, diff]
don't pull unavailable slots in with system and world 

This fix, in svn r4619, prevents the greedy system and world behavior (bug 4698) from pulling in unavailable slots.
Comment 12 Zac Medico gentoo-dev 2006-10-08 19:01:40 UTC
This has been released in 2.1.2_pre2-r7.
Comment 13 Duncan 2006-10-11 10:42:09 UTC
(In reply to comment #9)
> (In reply to comment #6)
> > Thanks, and on the CFLAGS, don't worry, if I can add to CFLAGS, I can just as
> > easily remove from CFLAGS, as routinely do before bug reports
> 
> Not even much funny... You are asking for your compile bugs to be marked
> INVALID on the fly, probably.

I think you misunderstood, perhaps as I missed the crucial bit.  If I can add to CFLAGS, I can just as easily remove from CFLAGS and try again before bugfiling.  That's what I meant, not that I remove CFLAGS as actually used from the reports filed, which is I think how you took it.  I'm not /that/ devious =8^), and yes, I agree, that wouldn't be funny at all!  =8^(