| 1 |
On 1/18/21 6:07 PM, Alec Warner wrote: |
| 2 |
> On Fri, Jan 15, 2021 at 6:47 PM Matt Turner <mattst88@g.o> wrote: |
| 3 |
>> |
| 4 |
>> This set is the upgradable packages for which the highest visible |
| 5 |
>> version has a different subslot than the currently installed version. |
| 6 |
>> |
| 7 |
>> The primary purpose of this feature is for use in catalyst builds. We |
| 8 |
>> update the "seed" stage3 before using it to build a new stage1. |
| 9 |
>> |
| 10 |
>> Updating the entire stage is expensive and unnecessary (since we're |
| 11 |
>> going to build the latest packages in stage1 and then rebuild everything |
| 12 |
>> in stage3). |
| 13 |
>> |
| 14 |
>> What we definitely do need to update in the original stage3 however, is |
| 15 |
>> any package that would trigger a subslot rebuild. |
| 16 |
>> |
| 17 |
>> For example: gcc links with libmpfr.so from dev-libs/mpfr. mpfr's SONAME |
| 18 |
>> changes from libmpfr.so.4 (SLOT="0/4") to libmpfr.so.6 (SLOT="0/6"). If |
| 19 |
>> the seed stage's dev-libs/mpfr is not updated before emerging gcc, gcc |
| 20 |
>> will link with libmpfr.so.4, but the latest version of dev-libs/mpfr |
| 21 |
>> will be built and libmpfr.so.6 included into the stage1. Since the old |
| 22 |
>> libmpfr.so.4 is not included in the stage1, gcc will not work, breaking |
| 23 |
>> subsequent stage builds. |
| 24 |
>> |
| 25 |
>> Our current options to update the seed are too large a hammer (e.g., |
| 26 |
>> "--update --deep --newuse @world" or "--update --deep --newuse |
| 27 |
>> --complete-graph --rebuild-if-new-ver gcc") and spend too much time |
| 28 |
>> updating seed stages for no gain beyond updating only packages for whom |
| 29 |
>> the subslot has changed. |
| 30 |
>> |
| 31 |
>> With this set, catalyst will likely use |
| 32 |
>> |
| 33 |
>> emerge @changed-subslot --ignore-built-slot-operator-deps y |
| 34 |
>> |
| 35 |
>> to update the seed stage. |
| 36 |
>> |
| 37 |
>> Thank you to Zac Medico for showing me how to do this. |
| 38 |
>> |
| 39 |
>> Bug: https://bugs.gentoo.org/739004 |
| 40 |
>> Signed-off-by: Matt Turner <mattst88@g.o> |
| 41 |
>> --- |
| 42 |
>> cnf/sets/portage.conf | 5 +++++ |
| 43 |
>> lib/portage/_sets/dbapi.py | 39 +++++++++++++++++++++++++++++++++++++- |
| 44 |
>> 2 files changed, 43 insertions(+), 1 deletion(-) |
| 45 |
>> |
| 46 |
>> diff --git a/cnf/sets/portage.conf b/cnf/sets/portage.conf |
| 47 |
>> index 22f0fa3a5..5651a9c53 100644 |
| 48 |
>> --- a/cnf/sets/portage.conf |
| 49 |
>> +++ b/cnf/sets/portage.conf |
| 50 |
>> @@ -84,6 +84,11 @@ exclude-files = /usr/bin/Xorg |
| 51 |
>> [rebuilt-binaries] |
| 52 |
>> class = portage.sets.dbapi.RebuiltBinaries |
| 53 |
>> |
| 54 |
>> +# Installed packages for which the subslot of the highest visible ebuild |
| 55 |
>> +# version is different than the currently installed version. |
| 56 |
>> +[changed-subslot] |
| 57 |
>> +class = portage.sets.dbapi.SubslotChangedSet |
| 58 |
>> + |
| 59 |
>> # Installed packages for which the highest visible ebuild |
| 60 |
>> # version is lower than the currently installed version. |
| 61 |
>> [downgrade] |
| 62 |
>> diff --git a/lib/portage/_sets/dbapi.py b/lib/portage/_sets/dbapi.py |
| 63 |
>> index 52367c4a6..46ba5c17d 100644 |
| 64 |
>> --- a/lib/portage/_sets/dbapi.py |
| 65 |
>> +++ b/lib/portage/_sets/dbapi.py |
| 66 |
>> @@ -15,7 +15,7 @@ from portage._sets import SetConfigError, get_boolean |
| 67 |
>> import portage |
| 68 |
>> |
| 69 |
>> __all__ = ["CategorySet", "ChangedDepsSet", "DowngradeSet", |
| 70 |
>> - "EverythingSet", "OwnerSet", "VariableSet"] |
| 71 |
>> + "EverythingSet", "OwnerSet", "SubslotChangedSet", "VariableSet"] |
| 72 |
>> |
| 73 |
>> class EverythingSet(PackageSet): |
| 74 |
>> _operations = ["merge"] |
| 75 |
>> @@ -167,6 +167,43 @@ class VariableSet(EverythingSet): |
| 76 |
>> |
| 77 |
>> singleBuilder = classmethod(singleBuilder) |
| 78 |
>> |
| 79 |
>> +class SubslotChangedSet(PackageSet): |
| 80 |
>> + |
| 81 |
>> + _operations = ["merge", "unmerge"] |
| 82 |
>> + |
| 83 |
>> + description = "Package set which contains all packages " + \ |
| 84 |
>> + "for which the subslot of the highest visible ebuild is " + \ |
| 85 |
>> + "different than the currently installed version." |
| 86 |
> |
| 87 |
> description = ("string1", |
| 88 |
> "string2", |
| 89 |
> "string3") |
| 90 |
> |
| 91 |
> vs concat + \ for line continuation? |
| 92 |
> |
| 93 |
> -A |
| 94 |
|
| 95 |
We also got flak on irc about the classmethod(singleBuilder) usage as |
| 96 |
opposed to @classmethod. This package set is formatted exactly like |
| 97 |
others in the file, so it's just a copy / paste thing. |
| 98 |
|
| 99 |
On the topic of python formatting, maybe we should use something like |
| 100 |
https://github.com/psf/black to automate it? |
| 101 |
-- |
| 102 |
Thanks, |
| 103 |
Zac |
Archives