"Just a reminder Opensource committers are voluntary contributors, you
can’t expect people to just jump on PRs to review as soon as they are
created. It takes time, maybe more than you expect for changes in area most
commiters aren’t experienced."

I totally get that and realize the pain. I am not expecting anyone to jump
on my ticket, but if the build bot isn't working for weeks on end that is
probably impacting everyone.For the last 6 years where I was at there are
these code scanning tools like veracode or whatever and if the artifact has
a problem they let you know, They want a path way for remediation.

Hadoop & friends is particularly "challenging". Like in this case, java8 is
broken. Java 8 :) I know java8 is an LTS version but geez. Aren't we on
Java 25 :)

Also I'm not trying to do "new cool stuff". It is more like "keep the
lights-on" stuff. For example:

https://issues.apache.org/jira/browse/HADOOP-19772

This plugin in use is deprecated. Other people are finding that based on
there (centos rocky etc) the statically linked protoc it is bringing in
from maven doesn't even work. I know it is not "oww ahh. so cool!" stuff,
but I think it is more than an existential threat.  I am finding something
like reload-4j inside hadoop projects. My impression of reload-4j is that
it was a stop gap solution to deal with that lo4j thing from *2 years ago.*
But it still is in the livy classpath because they are still trying to
support spark 2 and hadoop 3.2 or whatever, which is bringing along java 8
and protoc 2.5 etc.

I am down to do some dirty work, bumping versions running the tests. But if
I can't get any comms like "Holy crap awesome! I know we needed to fix that
plugin! It has been nagging me for 2 years" or like "we know the build is
broken maybe a week" or like "we love that old protobuf plugin how dare you
try to remove it" it is just hard to navigate this.

I have to express a cross cutting need. I got livy and it is sufficient
technically, However if i take the owasp plugin and scan it and it produces
400 CVE my risk owners want remediation. Then if there is too much friction
peope start looking for alternatives

looky here spark connect!
https://spark.apache.org/docs/latest/spark-connect-overview.html

Because it was probably just made with all the new hip stuff it probably
isnt dragging along java 8 and protobuf 2.5 people will switch to it.








On Fri, Jan 9, 2026 at 10:04 AM Ayush Saxena <[email protected]> wrote:

> ++ common-dev
>
> On Fri, 9 Jan 2026 at 20:21, Ayush Saxena <[email protected]> wrote:
> >
> > There is an Apache Slack, but very few people are active over there.
> >
> > I have pushed ur mail to common-dev@ so someone interested can help
> review.
> >
> > Just a reminder Opensource committers are voluntary contributors, you
> can’t expect people to just jump on PRs to review as soon as they are
> created. It takes time, maybe more than you expect for changes in area most
> commiters aren’t experienced.
> >
> > Anyway good luck contributing. If nobody comes to rescue. I will try to
> chase it maybe on some weekend when I have some bandwidth.
> >
> > -Ayush
> >
> >
> >
> > On 9 Jan 2026, at 7:08 PM, Edward Capriolo <[email protected]>
> wrote:
> >
> > 
> > Hello,
> > Is there an ETA?
> > Is there a chat room?
> >
> > Can we do this the old fashioned way? Review, run the test locally and
> merge?
> >
> > Again I really want to stress this: It is inefficient to communicate
> with the projects. It's like send an email, wait a day, send another email,
> similar issues with problem I am having with Livy, trying to understand why
> this is like that is anyone able to review something.
> >
> > Ed
> >
> > On Sun, Jan 4, 2026 at 11:03 AM HuaLong Zhang <[email protected]>
> wrote:
> >>
> >> Hi Edward,
> >> Sorry for the late reply.
> >> I’m already following up on this and have opened apache/hadoop#8162 to
> address it (switching the precommit Docker image default Java from JDK 8 to
> JDK 17).
> >> About the error: the build is being compiled with JDK 8 (javac 1.8),
> but the Maven compiler configuration passes --add-modules (e.g.,
> jdk.javadoc), which is a Java 9+ module-system option. JDK 8’s javac
> doesn’t understand that flag, so it fails with:
> >>
> >> Fatal error compiling: invalid flag: --add-modules
> >>
> >> Best,
> >> Hualong Zhang
> >> ________________________________
> >> From: Edward Capriolo <[email protected]>
> >> Sent: Thursday, January 1, 2026 16:01
> >> To: [email protected] <[email protected]>
> >> Cc: Ahmed Altaj <[email protected]>
> >> Subject: Re: Review https://issues.apache.org/jira/browse/HADOOP-19756
> >>
> >> I mean to ask, "Does hadoop-dev have a discord?"
> >>
> >>
> https://ci-hadoop.apache.org/job/hadoop-multibranch/job/PR-8156/lastFailedBuild/console
> >>
> >> https://issues.apache.org/jira/browse/HADOOP-19756?filter=-3
> >>
> >> I am aware it is the holidays, but it is very had to chase random
> contributors over email, to say "Hey I think the build bot is broken",
> would be easier in a chat do you all still do the IRC channel?
> >>
> >> Edward
> >>
> >> On Mon, Dec 29, 2025 at 2:10 PM Ahmed Altaj via user <
> [email protected]> wrote:
> >>
> >> Hi Edward
> >>
> >> Yahoo Mail: Search, organise, conquer
> >>
> >> On Mon, 29 Dec 2025 at 22:20, Edward Capriolo
> >> <[email protected]> wrote:
> >> Hey friends,
> >>
> >> I don't know if you all hang out on IRC or discord  to poke people to
> ask for reviews.  I checked the dev mailing list and see mostly auto-email.
> >>
> >> Consider this a poke. Thanks,
> >> Edward
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>

Reply via email to