| 1 |
On 7 June 2020 21:30:19 CEST, Rich Freeman <rich0@g.o> wrote: |
| 2 |
>On Sun, Jun 7, 2020 at 2:56 PM n952162 <n952162@×××.de> wrote: |
| 3 |
>> |
| 4 |
>> |
| 5 |
>> $ equery list \* | grep readline |
| 6 |
>> sys-libs/readline-7.0_p5-r1 |
| 7 |
>> |
| 8 |
>> But, given your answer about exclusivity/inclusivity in the other |
| 9 |
>thread, I guess this result is questionable... |
| 10 |
> |
| 11 |
>This is just showing what version you have installed, not what |
| 12 |
>versions are available. |
| 13 |
> |
| 14 |
>> |
| 15 |
>> The ebuild for bash-5.0_p17 has: |
| 16 |
>> |
| 17 |
>> READLINE_VER="8.0" |
| 18 |
>> |
| 19 |
>> The ebuilds for other the other users don't, I believe. |
| 20 |
> |
| 21 |
>So, first, this is just a random local variable and has no meaning in |
| 22 |
>and of itself. It is used in the dependency string which makes that |
| 23 |
>version of bash dependent on readline v8 specifically. Other packages |
| 24 |
>that don't list a version of readline will accept any version that |
| 25 |
>isn't masked/etc. So they're fine with v8. |
| 26 |
> |
| 27 |
>> |
| 28 |
>> The emerge that I used was this: |
| 29 |
>> |
| 30 |
>> emerge -auDv --verbose-conflicts --changed-use --keep-going |
| 31 |
>--with-bdeps=y --changed-deps --backtrack=100 @system |
| 32 |
> |
| 33 |
>Yeah, you might have to include the other packages that need readline |
| 34 |
>if portage complains. |
| 35 |
> |
| 36 |
>> |
| 37 |
>> However, I was just able to get it to build with this script: |
| 38 |
>> |
| 39 |
>> $ cat update-readline |
| 40 |
>> #!/usr/bin/env bash |
| 41 |
>> emerge -uUv $(cat <<-eof |
| 42 |
>> sys-libs/readline |
| 43 |
>> dev-db/postgresql |
| 44 |
>> sys-apps/gawk |
| 45 |
>> net-wireless/wpa_supplicant |
| 46 |
>> sys-fs/lvm2 |
| 47 |
>> dev-lang/python |
| 48 |
>> dev-lang/lua |
| 49 |
>> sci-visualization/gnuplot |
| 50 |
>> dev-db/postgresql |
| 51 |
>> app-text/hunspell |
| 52 |
>> sys-fs/udftools |
| 53 |
>> sys-block/parted |
| 54 |
>> x11-wm/fvwm |
| 55 |
>> net-misc/ntp |
| 56 |
>> sys-devel/gdb |
| 57 |
>> dev-db/postgresql |
| 58 |
>> sys-libs/gdbm |
| 59 |
>> net-mail/mailutils |
| 60 |
>> app-misc/rlwrap |
| 61 |
>> sys-devel/bc |
| 62 |
>> dev-libs/libxml2 |
| 63 |
>> net-dns/bind-tools |
| 64 |
>> eof |
| 65 |
>> ) |
| 66 |
>> |
| 67 |
> |
| 68 |
>That will probably work. Offhand I'm not sure if you need to add -1 / |
| 69 |
>--oneshot to that to prevent all that cruft from being added to your |
| 70 |
>world file. |
| 71 |
|
| 72 |
You need to add "-1" or "--oneshot". |
| 73 |
|
| 74 |
As this has been used, I would definitely expect the world-file to be full of this, causing issues with updates later. |
| 75 |
|
| 76 |
Also, by restricting to @system, any packages not in @system with a restriction on readline V8 will cause the mentioned problem. |
| 77 |
|
| 78 |
@system is, for me, a lasr resort, but I tend to move my world file away (rename) and put it back once @system is done and a depclean finished. This is usually only needed after not updating for a while and/or big changes in the tree. |
| 79 |
|
| 80 |
-- |
| 81 |
Joost |
| 82 |
-- |
| 83 |
Sent from my Android device with K-9 Mail. Please excuse my brevity. |
Archives