Skip to contentSkip to conversation

python3: upgrade 3.14.7 -> 3.14.8

PATCH In review Vijay Anusuri 5 participants 9 messages Oct 8 – Oct 21

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 python3: upgrade 3.14.7 -> 3.14.8

Release information:
* https://www.python.org/downloads/release/python-3148/
* Python 3.14.8 is an expedited security release and the eighth maintenance release of 3.14, containing around 354 bugfixes, build improvements and documentation changes from 142 contributors since 3.14.7.

Handles CVE-2026-19445 CVE-2026-19553 CVE-2026-19672 CVE-2026-15310 CVE-2026-15806 CVE-2026-17084

Changelog: https://docs.python.org/3/whatsnew/changelog.html#python-3-14-8-final

---
 .../python/{python3_3.14.7.bb => python3_3.14.8.bb}             | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
 rename meta/recipes-devtools/python/{python3_3.14.7.bb => python3_3.14.8.bb} (99%)
meta/recipes-devtools/python/python3_3.14.8.bb renamed from meta/recipes-devtools/python/python3_3.14.7.bb +1 -1
similarity index 99%
rename from meta/recipes-devtools/python/python3_3.14.7.bb
rename to meta/recipes-devtools/python/python3_3.14.8.bb
index f917edebd8..4663ee85cd 100644
@@ -30,7 +30,7 @@ SRC_URI:append:class-native = " \
            file://0001-Lib-sysconfig.py-use-prefix-value-from-build-configu.patch \
            "
 
-SRC_URI[sha256sum] = "3b48dac8fb59f62eaa67ac83c1eb12bda1b7a08406dd286e252c11a66be27f81"
+SRC_URI[sha256sum] = "c2215904f02b175596dc49351585104f4bc20341e1c47378b26a2c274360ce73"
 
 # exclude pre-releases for both python 2.x and 3.x
 UPSTREAM_CHECK_REGEX = "[Pp]ython-(?P<pver>\d+(\.\d+)+).tar"
Signature · 1 line
2.43.0
#1 lore raw
Quoted 11 lines · Vijay Anusuri via lists.openembedded.org

Release information:
* https://www.python.org/downloads/release/python-3148/
* Python 3.14.8 is an expedited security release and the eighth maintenance release of 3.14, containing around 354 bugfixes, build improvements and documentation changes from 142 contributors since 3.14.7.

Handles CVE-2026-19445 CVE-2026-19553 CVE-2026-19672 CVE-2026-15310 CVE-2026-15806 CVE-2026-17084

Changelog: https://docs.python.org/3/whatsnew/changelog.html#python-3-14-8-final

---

Hi Vijay,

Thanks for this upgrade, but we have failing ptests when building with musl:

WARNING: core-image-ptest-python3-1.0-r0 do_testimage: There were failing ptests.
Traceback (most recent call last):
File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/runtime/cases/ptest.py", line 27, in test_ptestrunner_expectfail
self.do_ptestrunner()
File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/runtime/cases/ptest.py", line 121, in do_ptestrunner
self.fail(failmsg)
AssertionError:
Failed ptests:
{'python3': ['test_append_limited_history', 'python3']}

https://autobuilder.yoctoproject.org/valkyrie/#/builders/109/builds/1353
https://autobuilder.yoctoproject.org/valkyrie/#/builders/110/builds/1341

This was recently discussed here:
https://lore.kernel.org/openembedded-core/[email protected]/

Are you willing to investigate this further? We need to understand if the test is failing because an issue in python code or because of a dependency we are using in oe-core.

Thanks,
Mathieu

Signature · 3 lines
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#2 lore raw
Quoted 41 lines · Mathieu Dubois-Briand via lists.openembedded.org
Quoted 11 lines · Vijay Anusuri via lists.openembedded.org

Release information:
* https://www.python.org/downloads/release/python-3148/
* Python 3.14.8 is an expedited security release and the eighth maintenance release of 3.14, containing around 354 bugfixes, build improvements and documentation changes from 142 contributors since 3.14.7.

Handles CVE-2026-19445 CVE-2026-19553 CVE-2026-19672 CVE-2026-15310 CVE-2026-15806 CVE-2026-17084

Changelog: https://docs.python.org/3/whatsnew/changelog.html#python-3-14-8-final

---

Hi Vijay,

Thanks for this upgrade, but we have failing ptests when building with musl:

WARNING: core-image-ptest-python3-1.0-r0 do_testimage: There were failing ptests.
Traceback (most recent call last):
File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
return func(*args, **kwargs)

          ^^^^^^^^^^^^^^^^^^^^^
 File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
   return func(*args, **kwargs)
          ^^^^^^^^^^^^^^^^^^^^^
 File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
   return func(*args, **kwargs)
          ^^^^^^^^^^^^^^^^^^^^^

File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/runtime/cases/ptest.py", line 27, in test_ptestrunner_expectfail
self.do_ptestrunner()
File "/srv/pokybuild/yocto-worker/qemux86-64-musl-ptest/build/layers/openembedded-core/meta/lib/oeqa/runtime/cases/ptest.py", line 121, in do_ptestrunner
self.fail(failmsg)
AssertionError:
Failed ptests:
{'python3': ['test_append_limited_history', 'python3']}

https://autobuilder.yoctoproject.org/valkyrie/#/builders/109/builds/1353
https://autobuilder.yoctoproject.org/valkyrie/#/builders/110/builds/1341

Quoted 6 lines

This was recently discussed here:
https://lore.kernel.org/openembedded-core/[email protected]/

Are you willing to investigate this further? We need to understand if the test is failing because an issue in python code or because of a dependency we are using in oe-core.

This is probably related to some libedit fixes that merged upstream, in an attempt to fix tests that were failing when libedit was being used:

https://github.com/python/cpython/commit/e87da800b6103fd890841688da48e41e211be7b9

Ross

#3 lore raw
Quoted 7 lines · Ross Burton via lists.openembedded.org
Quoted 3 lines

Are you willing to investigate this further? We need to understand if the test is failing because an issue in python code or because of a dependency we are using in oe-core.

This is probably related to some libedit fixes that merged upstream, in an attempt to fix tests that were failing when libedit was being used:

https://github.com/python/cpython/commit/e87da800b6103fd890841688da48e41e211be7b9

I replicated and filed https://github.com/python/cpython/issues/159424 upstream.

Ross

#4 lore raw

Hi Ross,

Thanks for investigating this and filing the upstream issue.

I was also able to reproduce this issue with Python 3.14.8 on our musl build.

Since the issue has been reported upstream, can we temporarily skip test_append_limited_history until a fix is available?

Thanks & Regards,

Vijay

Quoted 16 lines · Ross Burton
Quoted 5 lines · Ross Burton via lists.openembedded.org
Quoted 3 lines

Are you willing to investigate this further? We need to understand if the test is failing because an issue in python code or because of a dependency we are using in oe-core.

This is probably related to some libedit fixes that merged upstream, in

an attempt to fix tests that were failing when libedit was being used:

Quoted 2 lines

https://github.com/python/cpython/commit/e87da800b6103fd890841688da48e41e211be7b9

I replicated and filed https://github.com/python/cpython/issues/159424
upstream.

Ross

#5 lore raw
Quoted 6 lines · Vijay Anusuri

Hi Ross,
Thanks for investigating this and filing the upstream issue.

I was also able to reproduce this issue with Python 3.14.8 on our musl build.
Since the issue has been reported upstream, can we temporarily skip test_append_limited_history until a fix is available?

Only if we confirm that the test case is broken, and not the code in python. The point of the tests is to verify that Python works: if we disable failing tests, what’s the point?

Ross

#6 lore raw
In reply to Vijay Anusuri

Hello Vijay, Ross,

I reviewed the 3.14.8 upgrade. The downloaded source tarball’s SHA-256 matches the recipe, and the retained source patches apply without fuzz.

Following Ross’s upstream report [1], I inspected libedit’s history append path. append_history() opens the file with fopen(filename, "a"), and history_save_fp() writes the _HiStOrY_V2_ header whenever ftell() returns zero.

If the append stream initially reports offset zero on musl, this would cause another header to be appended to an existing history file. That is consistent with the reported result where the header appears as a history entry after truncation.

This suggests a history handling issue that needs investigation before deciding whether a test skip is appropriate. Could you check whether explicitly positioning the append stream at EOF before saving resolves the failure, and share the result in the upstream issue?

This analysis is based on source inspection. I have not independently reproduced the musl failure.

[1] https://github.com/python/cpython/issues/159424

Thanks,
Darsh

Quoted 42 lines · Vijay Anusuri

Release information:
* https://www.python.org/downloads/release/python-3148/
* Python 3.14.8 is an expedited security release and the eighth maintenance release of 3.14, containing around 354 bugfixes, build improvements and documentation changes from 142 contributors since 3.14.7.

Handles CVE-2026-19445 CVE-2026-19553 CVE-2026-19672 CVE-2026-15310 CVE-2026-15806 CVE-2026-17084

Changelog: https://docs.python.org/3/whatsnew/changelog.html#python-3-14-8-final

---
.../python/{python3_3.14.7.bb => python3_3.14.8.bb} | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)

rename meta/recipes-devtools/python/{python3_3.14.7.bb => python3_3.14.8.bb} (99%)

meta/recipes-devtools/python/python3_3.14.8.bb renamed from meta/recipes-devtools/python/python3_3.14.7.bb +0 -0
similarity index 99%
rename from meta/recipes-devtools/python/python3_3.14.7.bb
rename to meta/recipes-devtools/python/python3_3.14.8.bb
index f917edebd8..4663ee85cd 100644
@@ -30,7 +30,7 @@ SRC_URI:append:class-native = " \
 file://0001-Lib-sysconfig.py-use-prefix-value-from-build-configu.patch \

"

-SRC_URI[sha256sum] = "3b48dac8fb59f62eaa67ac83c1eb12bda1b7a08406dd286e252c11a66be27f81" +SRC_URI[sha256sum] = "c2215904f02b175596dc49351585104f4bc20341e1c47378b26a2c274360ce73"

# exclude pre-releases for both python 2.x and 3.x
UPSTREAM_CHECK_REGEX = "[Pp]ython-(?P<pver>\d+(\.\d+)+).tar"
--
2.43.0
#7 lore raw
Quoted 20 lines · On Tue, Oct 20, 2026 at 5:05 AM Darsh Kelaiya -X (dkelaiya - E INFOCHIPS…

Hello Vijay, Ross,

I reviewed the 3.14.8 upgrade. The downloaded source tarball’s SHA-256 matches the recipe, and the retained source patches apply without fuzz.

Following Ross’s upstream report [1], I inspected libedit’s history append path. append_history() opens the file with fopen(filename, "a"), and history_save_fp() writes the _HiStOrY_V2_ header whenever ftell() returns zero.

If the append stream initially reports offset zero on musl, this would cause another header to be appended to an existing history file. That is consistent with the reported result where the header appears as a history entry after truncation.

This suggests a history handling issue that needs investigation before deciding whether a test skip is appropriate. Could you check whether explicitly positioning the append stream at EOF before saving resolves the failure, and share the result in the upstream issue?

Quoted 5 lines

This analysis is based on source inspection. I have not independently reproduced the musl failure.

[1] https://github.com/python/cpython/issues/159424

I wonder why you did not look into the fix

https://github.com/python/cpython/pull/157165

which is also posted in the above
bug report. Does this fix the issue ?

Quoted 60 lines

Thanks,
Darsh

On Thu, Oct 8, 2026 at 02:31 PM, Vijay Anusuri wrote:

Release information:
* https://www.python.org/downloads/release/python-3148/
* Python 3.14.8 is an expedited security release and the eighth maintenance release of 3.14, containing around 354 bugfixes, build improvements and documentation changes from 142 contributors since 3.14.7.

Handles CVE-2026-19445 CVE-2026-19553 CVE-2026-19672 CVE-2026-15310 CVE-2026-15806 CVE-2026-17084

Changelog:
https://docs.python.org/3/whatsnew/changelog.html#python-3-14-8-final

---
.../python/{python3_3.14.7.bb => python3_3.14.8.bb} | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)

rename meta/recipes-devtools/python/{python3_3.14.7.bb => python3_3.14.8.bb} (99%)

meta/recipes-devtools/python/python3_3.14.8.bb renamed from meta/recipes-devtools/python/python3_3.14.7.bb +0 -0
similarity index 99%
rename from meta/recipes-devtools/python/python3_3.14.7.bb
rename to meta/recipes-devtools/python/python3_3.14.8.bb
index f917edebd8..4663ee85cd 100644
@@ -30,7 +30,7 @@ SRC_URI:append:class-native = " \
 file://0001-Lib-sysconfig.py-use-prefix-value-from-build-configu.patch \

"

-SRC_URI[sha256sum] = "3b48dac8fb59f62eaa67ac83c1eb12bda1b7a08406dd286e252c11a66be27f81" +SRC_URI[sha256sum] = "c2215904f02b175596dc49351585104f4bc20341e1c47378b26a2c274360ce73"

# exclude pre-releases for both python 2.x and 3.x
UPSTREAM_CHECK_REGEX = "[Pp]ython-(?P<pver>\d+(\.\d+)+).tar"
--
2.43.0

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group. View/Reply Online (#248274):
https://lists.openembedded.org/g/openembedded-core/message/248274
Mute This Topic: https://lists.openembedded.org/mt/121652032/1997914
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [ [email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

#8 lore raw

Hi Khem,

Quoted 113 lines · Khem Raj via lists.openembedded.org
Quoted 20 lines · On Tue, Oct 20, 2026 at 5:05 AM Darsh Kelaiya -X (dkelaiya - E INFOCHIPS…

Hello Vijay, Ross,

I reviewed the 3.14.8 upgrade. The downloaded source tarball’s SHA-256 matches the recipe, and the retained source patches apply without fuzz.

Following Ross’s upstream report [1], I inspected libedit’s history append path. append_history() opens the file with fopen(filename, "a"), and history_save_fp() writes the _HiStOrY_V2_ header whenever ftell() returns zero.

If the append stream initially reports offset zero on musl, this would cause another header to be appended to an existing history file. That is consistent with the reported result where the header appears as a history entry after truncation.

This suggests a history handling issue that needs investigation before deciding whether a test skip is appropriate. Could you check whether explicitly positioning the append stream at EOF before saving resolves the failure, and share the result in the upstream issue?

Quoted 5 lines

This analysis is based on source inspection. I have not independently reproduced the musl failure.

[1] https://github.com/python/cpython/issues/159424

I wonder why you did not look into the fix

https://github.com/python/cpython/pull/157165

which is also posted in the above
bug report. Does this fix the issue ?

The fix from PR #157165 is already included in Python 3.14.8, which we are upgrading to.

That commit fixes existing SKIPPED_TESTS exclusion for test_write_read_limited_history. This test is now passing with the latest Python version.

However, the newly introduced test test_append_limited_history is failing on musl builds.

Quoted 43 lines

Thanks,
Darsh

On Thu, Oct 8, 2026 at 02:31 PM, Vijay Anusuri wrote:

Release information:
* https://www.python.org/downloads/release/python-3148/
* Python 3.14.8 is an expedited security release and the eighth maintenance release of 3.14, containing around 354 bugfixes, build improvements and documentation changes from 142 contributors since 3.14.7.

Handles CVE-2026-19445 CVE-2026-19553 CVE-2026-19672 CVE-2026-15310 CVE-2026-15806 CVE-2026-17084

Changelog:
https://docs.python.org/3/whatsnew/changelog.html#python-3-14-8-final

---
.../python/{python3_3.14.7.bb => python3_3.14.8.bb} | 2 +-

1 file changed, 1 insertion(+), 1 deletion(-)

rename meta/recipes-devtools/python/{python3_3.14.7.bb => python3_3.14.8.bb} (99%)

meta/recipes-devtools/python/python3_3.14.8.bb renamed from meta/recipes-devtools/python/python3_3.14.7.bb +0 -0
similarity index 99%
rename from meta/recipes-devtools/python/python3_3.14.7.bb
rename to meta/recipes-devtools/python/python3_3.14.8.bb
index f917edebd8..4663ee85cd 100644
@@ -30,7 +30,7 @@ SRC_URI:append:class-native = " \
 file://0001-Lib-sysconfig.py-use-prefix-value-from-build-configu.patch \

"

-SRC_URI[sha256sum] = "3b48dac8fb59f62eaa67ac83c1eb12bda1b7a08406dd286e252c11a66be27f81" +SRC_URI[sha256sum] = "c2215904f02b175596dc49351585104f4bc20341e1c47378b26a2c274360ce73"

Quoted 9 lines
# exclude pre-releases for both python 2.x and 3.x
UPSTREAM_CHECK_REGEX = "[Pp]ython-(?P<pver>\d+(\.\d+)+).tar"
--
2.43.0

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group. View/Reply Online (#248304):
https://lists.openembedded.org/g/openembedded-core/message/248304
Mute This Topic: https://lists.openembedded.org/mt/121652032/7301997
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [ [email protected]]
-=-=-=-=-=-=-=-=-=-=-=-