https://bugzilla.redhat.com/show_bug.cgi?id=1866570
--- Comment #5 from Alex Scheel <ascheel(a)redhat.com> ---
Adam -- not sure I follow. My reading of what Alexander says (and our
experience with Debian) shows that the Dogtag proper doesn't have JDK 11
issues. Fedora's packaging of Dogtag's dependencies are what have Java 11
issues at the current moment. Most telling is that, while Debian also ships
Dogtag at a similar version, they don't have this particular issue, per
discussion with Timo on #freeipa. (He has one remaining Tomcat bug, but that
came up recently due to versions of bnd mismatched with what upstream Tomcat
was targeting).
If there are specific issues with Dogtag, these should be noted and resolved.
However, if the issues are in the dependencies, I think it is disingenuous to
call them Dogtag bugs, any more so than a bug in glibc shouldn't result in
several hundred bugs filed against every component that uses glibc.
Short of rewriting Dogtag to avoid one of our most heavily used libraries
(resteasy provides all our REST API infrastructure), there's no way to avoid
this on Dogtag's part. We should document the bug in the component the bug is
in, and provide a fix for that component. Even though Dogtag is the only
consumer of Resteasy in the published Fedora repositories.
If you -- or anyone else! -- would like to contribute to rewriting Dogtag to
avoid resteasy, consider this a warm invitation :-) I'll happily merge the code
upstream because it will save us many person-months of maintenance on a JBoss
package they don't support in Fedora and RHEL.
If you want to file a tracker bug blocking Fedora 33 because FreeIPA is broken,
shouldn't that tracker bug be filed against FreeIPA, as that is the
release-blocking component? Then we can file a separate bug against resteasy
and mark it blocking this bug, and so on, until all Dogtag dependencies are
fixed.
My 2c.
--
You are receiving this mail because:
You are on the CC list for the bug.