Let me just say that if other RPM packages actually depend on /usr/bin/python3, you'll get /usr/bin/python3. However if non RPM software you use depends on it, you are now in trouble. So what you expect is a reasonable thing to ask.

Thanks. I'll move this to the mailing list (a better place to discuss this).

What does yum --enablerepo=epel-testing update actually do then?

(I'll send it.)

I consider raising a new warning as a change of API, but I realize that it is probably disputable. Anyway, the update is in stable, probably to late to change anything, let's not bike shed then. I was surprised by this, and others might be as well. I'd consider an e-mail to dependent package maintainers as a reasonable thing to do.

Are we sure this is API compatible enough to go into stable Fedoras? See for example

I've added a build of python36 that has explicit conflicts with previous versions of python34.


Makes python3 build again (except test_asyncio).

BZ#1688284 Regression of OpenSSL 1.1.1b-1 in EVP_PBE_scrypt() with salt=NULL

Adding karma to my own fix, but I've verified the fix is working:

<mock-chroot> sh-5.0# rpm -q rpmlint
<mock-chroot> sh-5.0# rpmlint rpm*
6 packages and 0 specfiles checked; 4 errors, 57 warnings.
BZ#1668400 rpmlint needs locale data

For the record, this broke Python

python2-txsocksx-0: need python2-parsley on F30.

I forgot to add docs, so now the karma is reset. @pwalter if you could readd it, I would appreciate it. Thanks.

Got a scriptlet error, probably harmless:

ValueError: File context for /usr/lib/chromium-browser already defined
ValueError: File context for /usr/lib/chromium-browser/ already defined
ValueError: File context for /usr/lib/chrome-sandbox already defined

cura and cura-lulzbot seem to work fine. I haven't tested actual printing.

I can now build packages in rawhide mock. Thanks.

This update has been unpushed.

Thank You all.

BZ#1663357 Carla claims to provide but doesn't