| 1 |
On 7/25/2021 19:19, Sam James wrote: |
| 2 |
> |
| 3 |
> |
| 4 |
>> On 22 Jul 2021, at 16:00, Alice <alicef@g.o> wrote: |
| 5 |
>> |
| 6 |
>> |
| 7 |
>> Signed-off-by: Alice Ferrazzi <alicef@g.o> |
| 8 |
>> --- |
| 9 |
>> eclass/kernel-2.eclass | 13 +++++++++---- |
| 10 |
>> 1 file changed, 9 insertions(+), 4 deletions(-) |
| 11 |
> |
| 12 |
> Alice, thanks for taking the initiative to get this done and drop Python 2.7 here. |
| 13 |
> |
| 14 |
> It's much appreciated! |
| 15 |
> |
| 16 |
>> |
| 17 |
>> diff --git a/eclass/kernel-2.eclass b/eclass/kernel-2.eclass |
| 18 |
>> index f94dd9c..05f8161 100644 |
| 19 |
>> --- a/eclass/kernel-2.eclass |
| 20 |
>> +++ b/eclass/kernel-2.eclass |
| 21 |
>> @@ -605,7 +605,7 @@ if [[ ${ETYPE} == sources ]]; then |
| 22 |
>> kernel_is le 2 6 ${DEBLOB_MAX_VERSION} && \ |
| 23 |
>> K_DEBLOB_AVAILABLE=1 |
| 24 |
>> if [[ ${K_DEBLOB_AVAILABLE} == 1 ]]; then |
| 25 |
>> - PYTHON_COMPAT=( python2_7 ) |
| 26 |
>> + PYTHON_COMPAT=( python3_{7..10} ) |
| 27 |
> |
| 28 |
> I think others already said this, but Python 3.7 is a noop (eclasses ignore it), so just drop it. |
| 29 |
> |
| 30 |
>> |
| 31 |
>> inherit python-any-r1 |
| 32 |
>> |
| 33 |
>> @@ -1489,9 +1489,14 @@ kernel-2_src_compile() { |
| 34 |
>> [[ ${ETYPE} == headers ]] && compile_headers |
| 35 |
>> |
| 36 |
>> if [[ ${K_DEBLOB_AVAILABLE} == 1 ]] && use deblob; then |
| 37 |
>> - einfo ">>> Running deblob script ..." |
| 38 |
>> - python_setup |
| 39 |
>> - sh "${T}/${DEBLOB_A}" --force || die "Deblob script failed to run!!!" |
| 40 |
>> + # deblob less than 5.10 require python 2.7 |
| 41 |
>> + if kernel_is lt 5 10; then |
| 42 |
>> + ewarn "we don't support deblob for kernel less then 5.10" |
| 43 |
>> + else |
| 44 |
>> + einfo ">>> Running deblob script ..." |
| 45 |
>> + python_setup |
| 46 |
>> + sh "${T}/${DEBLOB_A}" --force || die "Deblob script failed to run!!!" |
| 47 |
>> + fi |
| 48 |
>> fi |
| 49 |
>> } |
| 50 |
>> |
| 51 |
> |
| 52 |
> I'd still prefer it if we just didn't provide the USE flag on irrelevant kernels (it's too confusing) but if you plan on removing this |
| 53 |
> message at some point, I guess we can keep it? |
| 54 |
|
| 55 |
I think the message text for the ewarn case could be a bit better. How |
| 56 |
about instead of: |
| 57 |
|
| 58 |
"we don't support deblob for kernel less then 5.10" |
| 59 |
|
| 60 |
Something like: |
| 61 |
|
| 62 |
"Deblob support requires a Linux kernel >=5.10" |
| 63 |
|
| 64 |
Is used instead? |
| 65 |
|
| 66 |
Also, the die message text can drop the triple exclamation marks. One |
| 67 |
exclamation mark is enough. |
| 68 |
|
| 69 |
And someone correct me if wrong, as I am reading this thread late, but |
| 70 |
deblobbing is disabled by setting K_DEBLOB_AVAILABLE=0 in a kernel-2 |
| 71 |
inheritor ebuild, right? mips-sources can't support deblobbing, as several |
| 72 |
of the SGI machines (IP27, IP30) are literally unbootable off of internal |
| 73 |
drive bays if specific Qlogic firmware blobs don't get included in the |
| 74 |
kernel image. |
| 75 |
|
| 76 |
-- |
| 77 |
Joshua Kinard |
| 78 |
Gentoo/MIPS |
| 79 |
kumba@g.o |
| 80 |
rsa6144/5C63F4E3F5C6C943 2015-04-27 |
| 81 |
177C 1972 1FB8 F254 BAD0 3E72 5C63 F4E3 F5C6 C943 |
| 82 |
|
| 83 |
"The past tempts us, the present confuses us, the future frightens us. And |
| 84 |
our lives slip away, moment by moment, lost in that vast, terrible in-between." |
| 85 |
|
| 86 |
--Emperor Turhan, Centauri Republic |
Archives