On 17/05/2021 20:58, Neal Gompa wrote:
On Mon, May 17, 2021 at 3:57 PM Mike McGrath mmcgrath@redhat.com wrote:
On Mon, May 17, 2021 at 2:48 PM Phil Perry pperry@elrepo.org wrote:
On 17/05/2021 20:46, Phil Perry wrote:
On 17/05/2021 20:30, Johnny Hughes wrote:
On 5/17/21 1:46 PM, Mike McGrath wrote:
Snipping content as we seem to have reached some mailing list limit
This is mostly because we also don't ship it in RHEL and we don't ship it there because we don't want to be on the hook to support it (even RHEL has a budget and limited resources).
As far as CentOS Stream, the promise is that if it runs in CentOS Stream today it should run in the next version of RHEL. By including content in Stream that we don't include in RHEL, suddenly that promise is broken if someone accidentally uses it assuming it's in RHEL.
I'm not saying there aren't ways to work through these issues, but I wanted to give some ideas as to the thought process that got us here.
-Mike
Mike,
That is all well and good .. but it is your guys (EPEL and Red Hat SIGs) that need this Development content to be in CentOS Stream and the CentOS Community Build System to build things. (I can build things as this content is already in my Koji buildroot).
And it is me, this list, and the CentOS Stream group that keep getting asked (by EPEL and the RH SIGs) why this open source stuff can not be in CentOS Linux, CentOS Stream and the CBS.
Surely we can set up a non RHEL released repo that your guys can use to build the things that they want.
Thanks, Johnny Hughes
But from what Mike just said, EPEL/SIGs should be building against those missing -devel packages on Stream because anything built on Stream must also run on RHEL?
Sorry, typo - EPEL/SIGs should NOT be building against those missing -devel packages...
I'm always confused by this because the basis of CRB *was* supposed to be what was required to build EPEL. I'm not sure if EPEL has grown since we started or if we just missed the mark with CRB.
Unfortunately, I think it's pretty much the latter. :(
Perhaps this is a policy decision that can be reverted in RHEL 9 as it clearly isn't working very well :-)