| 1 |
On 2/16/17 2:19 PM, Walter Dnes wrote: |
| 2 |
> I'm installing uclibc-ng Gentoo on an old Atom netbook. I saw the |
| 3 |
> news item about uclibc-ng upgrade and wanted to do the switchover during |
| 4 |
> the install. This way "emerge -e @world" has to do "only" 153 packages |
| 5 |
> ...3 times ...over ten hours each time on the ancient netbook. I |
| 6 |
> finished step 5, and need to do only... |
| 7 |
> |
| 8 |
>> 6. Finally update uclibc-ng to the latest |
| 9 |
>> |
| 10 |
>> emerge =sys-libs/uclibc-ng-1.0.22 |
| 11 |
>> |
| 12 |
>> 7. For good measure, rebuild the entire system |
| 13 |
>> |
| 14 |
>> emerge â<80><94>e @world |
| 15 |
> uclibc-ng-1.0.22 is "masked by missing keywords"... |
| 16 |
> |
| 17 |
> # grep KEYWORDS /usr/portage/sys-libs/uclibc-ng/uclibc-ng-1.0.22.ebuild |
| 18 |
> #KEYWORDS="-* ~amd64 ~arm ~mips ~ppc ~x86" |
| 19 |
> KEYWORDS="" |
| 20 |
> |
| 21 |
> This is usually done for a reason. Are there any showstoppers with |
| 22 |
> uclibc-ng-1.0.22 and should I wait at uclibc-ng-1.0.20? Or is it OK to |
| 23 |
> tweak the ebuild and finish the upgrade? |
| 24 |
> |
| 25 |
You need to be careful because obstack support has been dropped in |
| 26 |
uclibc-ng-1.0.22. This means that if you directly upgrade to 1.0.22 |
| 27 |
without first rebuilding your systems (including the toolchain) without |
| 28 |
obstack support, you will have a severely broken system. The news item |
| 29 |
specifies the steps by which you can rebuild @world without depending on |
| 30 |
uclibc-ng to provide obstack, and then upgrading uclibc-ng. |
| 31 |
|
| 32 |
I've still got things masked because I'm giving people some time. I'm |
| 33 |
going to stabilize 1.0.20 today-ish and then 1.0.22 later this month. |
| 34 |
|
| 35 |
-- |
| 36 |
Anthony G. Basile, Ph.D. |
| 37 |
Gentoo Linux Developer [Hardened] |
| 38 |
E-Mail : basile@××××××××××.net |
| 39 |
GnuPG FP : 1FED FAD9 D82C 52A5 3BAB DC79 9384 FA6E F52D 4BBA |
| 40 |
GnuPG ID : F52D4BBA |
Archives