root@spawner etc # ls -al /etc/make.profile lrwxrwxrwx 1 root root 37 Apr 7 10:44 /etc/make.profile -> /usr/portage/profiles/selinux-x86-1.4 root@spawner etc # emerge rsync [..] total size is 62887805 speedup is 30.43 >>> Updating Portage cache... ...done! Traceback (most recent call last): File "/usr/bin/emerge", line 2250, in ? reload(portage) File "/usr/lib/portage/pym/portage.py", line 6128, in ? settings.regenerate() # XXX: Regenerate use after we get a vartree -- GLOBAL File "/usr/lib/portage/pym/portage.py", line 1392, in regenerate self.configdict["auto"]["USE"]=autouse(db[root]["vartree"],use_cache=use_cache) File "/usr/lib/portage/pym/portage.py", line 1121, in autouse myresult=dep_check(mydep,myvartree.dbapi,None,use="no",use_cache=use_cache) File "/usr/lib/portage/pym/portage.py", line 3309, in dep_check mylist=flatten(dep_listcleanup(dep_zapdeps(mysplit,mysplit2))) File "/usr/lib/portage/pym/portage.py", line 3052, in dep_zapdeps myresult=dep_zapdeps(unreduced[x],reduced[x]) File "/usr/lib/portage/pym/portage.py", line 3038, in dep_zapdeps elif myportapi.match(x): AttributeError: 'NoneType' object has no attribute 'match' root@@spawner portage ## emerge rsync Traceback (most recent call last): File "/usr/bin/emerge", line 14, in ? import portage File "/usr/lib/portage/pym/portage.py", line 6128, in ? settings.regenerate() # XXX: Regenerate use after we get a vartree -- GLOBAL File "/usr/lib/portage/pym/portage.py", line 1392, in regenerate self.configdict["auto"]["USE"]=autouse(db[root]["vartree"],use_cache=use_cache) File "/usr/lib/portage/pym/portage.py", line 1121, in autouse myresult=dep_check(mydep,myvartree.dbapi,None,use="no",use_cache=use_cache) File "/usr/lib/portage/pym/portage.py", line 3309, in dep_check mylist=flatten(dep_listcleanup(dep_zapdeps(mysplit,mysplit2))) File "/usr/lib/portage/pym/portage.py", line 3052, in dep_zapdeps myresult=dep_zapdeps(unreduced[x],reduced[x]) File "/usr/lib/portage/pym/portage.py", line 3038, in dep_zapdeps elif myportapi.match(x): AttributeError: 'NoneType' object has no attribute 'match' root@spawner etc ## date Wed Apr 7 14:09:17 EEST 2004 emerge can be used again only after rm -rf /usr/portage/profiles/selinux-x86-1.4 which happens to be my /etc/make.profile
Something is wrong with "virtuals" file. I've copied one from profiles/default-x86-1.4 to profiles/default-x86-2004.0 and portage seems to work now.
someone has been adding 'x11-base/xorg-x11' to all virtuals. without them, portage works again. (14:48:51) peter: root@@spawner selinux-x86-1.4 ## diff -u virtuals virtuals.old --- virtuals 2004-04-07 14:46:07.000000000 +0300 +++ virtuals.old 2004-04-07 12:17:15.000000000 +0300 @@ -12,7 +12,7 @@ virtual/flim app-emacs/flim virtual/ghc dev-lang/ghc-bin virtual/glibc sys-libs/glibc -virtual/glu x11-base/xfree +virtual/glu x11-base/xfree x11-base/xorg-x11 virtual/glut media-libs/glut virtual/imap-c-client net-libs/c-client virtual/imapUW net-mail/uw-imap @@ -29,7 +29,7 @@ virtual/mda net-mail/procmail virtual/modutils sys-apps/modutils virtual/mta net-mail/ssmtp -virtual/opengl x11-base/xfree +virtual/opengl x11-base/xfree x11-base/xorg-x11 virtual/os-headers sys-kernel/linux-headers virtual/php dev-php/mod_php virtual/python dev-lang/python @@ -40,9 +40,9 @@ virtual/sylpheed net-mail/sylpheed virtual/tetex app-text/tetex virtual/textbrowser net-www/links -virtual/x11 x11-base/xfree +virtual/x11 x11-base/xfree x11-base/xorg-x11 virtual/xemacs app-editors/xemacs -virtual/xft x11-base/xfree +virtual/xft x11-base/xfree x11-base/xorg-x11 x11-libs/xaw x11-libs/Xaw3d virtual/tftp net-misc/tftp-hpa virtual/gzip app-arch/gzip
x11-base/xorg-x11 was added as secondary x11 virtual, but the package itself wasn't commited yet. just resync in some minutes.
looks like removing all occurences of "x11-base/xorg-x11" from /etc/make.profile/virtuals is a workaround...
and besides that it's not that cool that a "broken" virtuals file can knock out the whole portage system :-((
It should be fixed in 2.0.50-r3 or later. Is there person who encounter this problem in the versions?
root@@spawner / ## epm -qa |grep portage portage-2.0.50-r3 i already had 2.0.50-r3 when it happened. also 2.0.50-r1 brakes in the same way.
I thought it was a portage *tree* problem and not a portage *program* problem?!
It's portage problem because portage shouldn't be broken even if there is problem in portage tree. We are fixing it now...
well, yes, it's a portage problem from that point of view, but the *reason* of the impact was a "broken" file :)
can someone please fix the portage ??? rsync breaks in the same way 5 hours after the bug has been posted. a portage that has been updated 5 minutes from Server Address : 217.72.114.99 Hardware : PIV 2.66 Ghz, 1024 Megs of RAM Connectivity : 100 Mbit/s user limit : 100 Location : Munich, Germany looks like this: Traceback (most recent call last): File "/usr/bin/emerge", line 2250, in ? reload(portage) File "/usr/lib/portage/pym/portage.py", line 6128, in ? settings.regenerate() # XXX: Regenerate use after we get a vartree -- GLOBAL File "/usr/lib/portage/pym/portage.py", line 1392, in regenerate self.configdict["auto"]["USE"]=autouse(db[root]["vartree"],use_cache=use_cache) File "/usr/lib/portage/pym/portage.py", line 1121, in autouse myresult=dep_check(mydep,myvartree.dbapi,None,use="no",use_cache=use_cache) File "/usr/lib/portage/pym/portage.py", line 3309, in dep_check mylist=flatten(dep_listcleanup(dep_zapdeps(mysplit,mysplit2))) File "/usr/lib/portage/pym/portage.py", line 3052, in dep_zapdeps myresult=dep_zapdeps(unreduced[x],reduced[x]) File "/usr/lib/portage/pym/portage.py", line 3038, in dep_zapdeps elif myportapi.match(x): AttributeError: 'NoneType' object has no attribute 'match' zak selinux-x86-1.4 # cat /etc/make.profile/virtuals |grep xorg virtual/x11 x11-base/xfree x11-base/xorg-x11 virtual/opengl x11-base/xfree x11-base/xorg-x11 virtual/glu x11-base/xfree x11-base/xorg-x11 virtual/xft x11-base/xfree x11-base/xorg-x11 and xorg-x11-6.7.0.ebuild is ~x86 or do the mirrors base themselves on emerge functionalities and they cannot get new portage trees? BTW why in god's name do i need X in selinux anyway ?!?!? the sex (secure X) project from the nsa is far from being done.
*** Bug 47105 has been marked as a duplicate of this bug. ***
portage-2.0.50-r3 is a security update of -r1 and does not contain this fix. 'emerge sync' will remedy this problem if you are still experiencing it.
I'd rather recommend waiting for a few hours until emerge sync, since most mirrors are apparently not updated yet, so it's still sed -i s#x11-base/xorg-x11## /etc/make.profile/virtuals
carpaski: 'emerge sync' will end in error (see my first posting). the only way to succesfully do ANYTHING with `emerge` or `ebuild` is to remove all xorg-x11 strings from the virtuals file. please also have in mind that the only xorg-x11 ebuild (xorg-x11-6.7.0.ebuild) is ~x86, so it's kinda masked for my non-devel x86 servers.
*** Bug 47070 has been marked as a duplicate of this bug. ***
Fixed in CVS.
I rsync'd three machines - one had ~x86 set. Of the three, the only one that did not exhibit this behavior was the one with ~x86 set.
*** Bug 47108 has been marked as a duplicate of this bug. ***
Well. No can do. emerge sync won't work. Seems we have a -bad- flaw here where a broken tree will render the whole system unusable. :-/
Same issue here (can't sync). For clarification, the errors that appear are the same as the errors in first post.
Same here. After adjusting /etc/make.profile/virtuals manually it works, but during "emerge sync" it breaks again: sys-libs/libtermcap-compat/libtermcap-compat-2.0.8-r1.ebuild 1698 100% 0.00kB/s 0:00:00 x11-base/xorg-x11/ChangeLog 712 100% 0.00kB/s 0:00:00 x11-base/xorg-x11/Manifest 239 100% 0.00kB/s 0:00:00 x11-base/xorg-x11/xorg-x11-6.7.0.ebuild 36967 100% 0.00kB/s 0:00:00 deleting profiles/default-x86-1.4/virtuals.old 200 files.... Number of files: 80537 Number of files transferred: 132 Total file size: 62931068 bytes Total transferred file size: 362701 bytes Literal data: 362701 bytes Matched data: 0 bytes File list size: 1821852 Total bytes written: 2297 Total bytes read: 1963247 wrote 2297 bytes read 1963247 bytes 49760.61 bytes/sec total size is 62931068 speedup is 32.02 >>> Updating Portage cache... ...done! Traceback (most recent call last): File "/usr/bin/emerge", line 2250, in ? reload(portage) File "/usr/lib/portage/pym/portage.py", line 6128, in ? settings.regenerate() # XXX: Regenerate use after we get a vartree -- GLOBAL File "/usr/lib/portage/pym/portage.py", line 1392, in regenerate self.configdict["auto"]["USE"]=autouse(db[root]["vartree"],use_cache=use_cache) File "/usr/lib/portage/pym/portage.py", line 1121, in autouse myresult=dep_check(mydep,myvartree.dbapi,None,use="no",use_cache=use_cache) File "/usr/lib/portage/pym/portage.py", line 3309, in dep_check mylist=flatten(dep_listcleanup(dep_zapdeps(mysplit,mysplit2))) File "/usr/lib/portage/pym/portage.py", line 3052, in dep_zapdeps myresult=dep_zapdeps(unreduced[x],reduced[x]) File "/usr/lib/portage/pym/portage.py", line 3038, in dep_zapdeps elif myportapi.match(x): AttributeError: 'NoneType' object has no attribute 'match' aero make.profile #
A workaround seems to be to apply the following patch: ------------------------------------------------------------------------ --- /usr/lib/portage/pym/ORIG-portage.py 2004-04-07 20:24:32.223100496 +0200 +++ /usr/lib/portage/pym/portage.py 2004-04-07 20:43:28.690331216 +0200 @@ -3031,11 +3031,11 @@ if (type(x)==types.ListType): match=1 for y in x: - if not myportapi.match(y): + if myportapi and not myportapi.match(y): match=0 if match: return x - elif myportapi.match(x): + elif myportapi and myportapi.match(x): return x #none of the not masked pkg, use the first one ------------------------------------------------------------------------ This makes emerge not crash when the portage tree is broken in this particular way.
I have experienced this problem as well. When applying your patch, it starts to rebuild the cache and then just sits there doing nothing forever... I now have an extremely broken portage and no portage-rescue is available for x86 anymore?
I have this prolem trying to do a brand new install from the x86 2004.0 boot CD. You can't even complete the initial sync. After I get the tree, it crashes. I tried the patch in comment 23, then tried to run emerge again, getting the comment 24 result (ie it updates, then waits.... seemingly forever). The only way I could make it work was by wiping the files that the stage2 tarball installed, re-extract them all, apply the patch, then run a emerge sync. After that, I seem to have a working portage (which is now updating).
I had the exact same problem. What I needed to do was : - Apply the patch from comment 23 before doing anything else besides unpacking the stage1 tarball. - emerge sync - cd /usr/portage - scripts/bootstrap.sh -f And then (it is getting a little tricky) I had to apply patch from commant 23 on the downloaded portage package, modify the MD5 digest corresponding to the package so that it doesn't get downloaded again. That last part was needed b/c bootstrap.sh updates portage and thus undoes the patch applied to portage.py rendering it unusable. I hope this helps.
*** Bug 47223 has been marked as a duplicate of this bug. ***
This bug has been fixed in CVS and -r2/-r4(unstable version) carpaski: we should release it as possible as we can.
thank you for the bugfix
