On 1/14/22 10:33, Troy Dawson wrote:
On Fri, Jan 14, 2022 at 8:13 AM Stephen John Smoogen <smooge@xxxxxxxxx <mailto:smooge@xxxxxxxxx>> wrote:On Fri, 14 Jan 2022 at 10:57, Stephen John Smoogen <smooge@xxxxxxxxx <mailto:smooge@xxxxxxxxx>> wrote: I mirrored the source rpms down and did the following for 8 and 9-stream. ``` $ for i in AppStream BaseOS PowerTools; do echo $i; find ./$i -type f -name "*src.rpm" | xargs rpm --nosignature --qf='%{NAME}\n' -qp > /tmp/a-$i; sort -o /tmp/a-$i -u /tmp/a-$i; done $ sort -o /tmp/a -u /tmp/a-* ; sort -o /tmp/b -u /tmp/b-* $ wc -l /tmp/a* /tmp/b* 2652 /tmp/a 1740 /tmp/a-AppStream 536 /tmp/a-BaseOS 503 /tmp/a-PowerTools 2273 /tmp/b 1620 /tmp/b-AppStream 399 /tmp/b-BaseOS 295 /tmp/b-CRB $ comm -1 -2 /tmp/a /tmp/b | wc -l 2090 $ comm -1 -3 /tmp/a /tmp/b | wc -l 183 $ comm -2 -3 /tmp/a /tmp/b | wc -l 562 ``` So 183 packages were added to 9 that weren't in 8 and 562 packages were 'removed'. Some of those are obsolete packages like python2, python36,python38, gcc-toolset-9, gcc-toolset-10,autoconf213. Others are module things which aren't shipped already.The following statement was wrong. Some subset of that 500 may be built and could go into CRB, but that would require mirroring the CentOS Stream koji which I didn't do. That leaves about 500 source packages which aren't even built internally so aren't going into CRB.I was gathering each of the names of the binary and source packages directly from the repos for my "Will It Build", so I did a few tweaks and got these numbers. I feel they are quite accurate.CentOS Stream 8: AppStream: 4553 BaseOS: 1715 CRB: 1614 Total Source RPMS: 2260 CentOS Stream 9: AppStream: 5225 BaseOS: 1130 CRB: 1370 Total Source RPMS: 2252So, there is a drop of 250 packages in CRB from RHEL8 to RHEL9. But beyond that, things are quite close.Troy
First off, I apologize for: - injecting bad data into this discussion- making assumptions about packages being in the RHEL9 buildroot that were not true
- using a subject that didn't really reflect what my real concern wasSo, I really don't care if RHEL overall is shrinking or not. My concern was over the possible increase in 'buildroot-only' packages - which seems may not be any worse than it is with RHEL8 (although that's annoying enough - more on that in another response).
Thanks Troy and Smooge for your data. But I guess what I would be really interested in finding out (if possible) is the number of buildroot-only packages in the two distros.
I've done some poking, and this is what I've come up with for *new* missing -devel packages in CS9 with their approximate number of users in rawhide:
389-ds-base-devel 2 accel-config-devel 0 anthy-unicode-devel 3 atlas-corei2-devel 0 avahi-glib-devel 1 avahi-gobject-devel 4 avahi-ui-devel 1 bind-devel 3 blis-devel 2 Box2D-devel 1 cjose-devel 1 clutter-devel 9 clutter-gst3-devel 1 clutter-gtk-devel 8 cogl-devel 0 compat-lua-devel 26 cppunit-devel 35 double-conversion-devel 9 dpdk-devel 1 emacs-devel 1 evince-devel 3 fdk-aac-free-devel 4 flashrom-devel 1 flexiblas-devel 33 freeglut-devel 68 fstrm-devel 4 glusterfs-api-devel 8 glusterfs-devel 2 gnome-calculator-devel 0 gnome-menus-devel 1 gnome-software-devel 0 gsl-devel 51 gtk4-devel-tools 0 gtksourceview4-devel 6 hexchat-devel 1 hidapi-devel 11 ibus-anthy-devel 0 ibus-table-devel 7 inih-devel 3 iscsi-initiator-utils-devel 1 isns-utils-devel 1 kernel-rt-debug-devel-matched 0 kernel-rt-devel-matched 0 ldns-devel 8 libass-devel 1 libblockdev-crypto-devel 1 libblockdev-devel 1 libblockdev-fs-devel 1 libblockdev-loop-devel 1 libblockdev-lvm-devel 1 libblockdev-mdraid-devel 1 libblockdev-part-devel 1 libblockdev-swap-devel 2 libblockdev-utils-devel 0 libbytesize-devel 5 libcbor-devel 0 libcephfs-devel 2 libcephsqlite-devel 0 libdazzle-devel 1 libeconf-devel 1 libell-devel 3 libepubgen-devel 0 libev-devel 22 libev-libevent-devel 0 libfido2-devel 2 libfl-devel 3 libfoma-devel 0 libhandy-devel 2 libldac-devel 0 libmng-devel 12 libmpeg2-devel 2 libmtp-devel 7 libotr-devel 8 libqrtr-glib-devel 2 libradosstriper-devel 1 libshaderc-devel 1 libslirp-devel 3 libsmartcols-devel 1 libsmi-devel 1 libstoragemgmt-devel 1 libtpms-devel 1 liburing-devel 6 libverto-libev-devel 0 libwpe-devel 0 libXpresent-devel 2 libzip-devel 16 lldpd-devel 0 lpsolve-devel 1 luajit-devel 11 make-devel 5 mecab-devel 7 mesa-vulkan-devel 2 minizip-compat-devel 6 mptcpd-devel 0 nmstate-devel 0 nodejs-devel 19 ocaml-camomile-devel 5 ocaml-csexp-devel 3 ocaml-dune-devel 17 physfs-devel 17 postgresql-upgrade-devel 1 pybind11-devel 19 qgpgme-devel 3 rados-objclass-devel 0 rapidjson-devel 18 sid-base-libs-devel 0 sid-iface-libs-devel 0 sid-log-libs-devel 0 sid-resource-libs-devel 0 speech-tools-libs-devel 1 suitesparse64-devel 3 suitesparse64_-devel 1 swtpm-devel 0 sysprof-devel 1 tesseract-devel 5 tix-devel 8 uchardet-devel 4 umockdev-devel 2 unbound-devel 8 v8-devel 2 vm-dump-metrics-devel 0 volume_key-devel 1 web-assets-devel 66 wireplumber-devel 0 wpebackend-fdo-devel 1 xkbcomp-devel 0 xorg-x11-drv-evdev-devel 0That's a lot. Now, many of these are pretty obscure and do not have any users - but I'm also seeing a number that are going to hit me and the scitech sig:
blis-devel 2 flexiblas-devel 33 gsl-devel 51 suitesparse64-devel 3 There are some other big ones as well. -- Orion Poplawski he/him/his - surely the least important thing about me Manager of NWRA Technical Systems 720-772-5637 NWRA, Boulder/CoRA Office FAX: 303-415-9702 3380 Mitchell Lane orion@xxxxxxxx Boulder, CO 80301 https://www.nwra.com/
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ epel-devel mailing list -- epel-devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to epel-devel-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/epel-devel@xxxxxxxxxxxxxxxxxxxxxxx Do not reply to spam on the list, report it: https://pagure.io/fedora-infrastructure
