Hi Phil,

On 13.11.2025 23:12, Phil Steitz wrote:
> I agree with the sentiment here, but I also understand Vladmir's position.
> We should formally EOL lang2 (and *many* other n-k versions of Commons
> components) unless we are willing to backport security fixes.


+1 on declaring Commons Lang 2.x EOL. We should have done this a long
time ago.

Without an *explicit* declaration, user expectations vary: some assume
any version not binary-compatible with the latest is unsupported, while
others expect a 14-year-old release to still be maintained. Since
`commons-logging` remained “supported” despite 9 years without a
release, those expectations aren’t unfounded. Personally, I prefer to
ask the maintainers [1] rather than make assumptions.

In the near future, Apache Trusted Releases will let us use the upcoming
Common Lifecycle Enumeration ECMA standard (see [1] and [2]), but for
now we need a vote and an *explicit* website update.

A brief ECMA note: the standard developed by TC54-TG4 (OSS
Sustainability) will define “support” levels [4] for projects and
releases. It urgently needs experienced contributors, including a
“co-convenor.” Since the ASF has been an ECMA member since June,
volunteers can contact me at [email protected] (or just join the OWASP
CycloneDX Slack).


> While the one dependency trace he posted may not be "real" and 99%
> of others may "miss" the CVE, it is not practical for users to
> validate these things and build tools are going to kick them out.

We now have a new set of tools [5] that automate the kind of analysis
Emmanuel showed in this thread, pre-computing call graphs of all
dependencies.

I will start running this for Solr 9.10.0’s dependencies today, and I
can also do it for gradle-nexus-staging-plugin so Vladimir can dismiss
that false positive.


> I will start another thread on the general topic, but I think we
> should provide a backport patch for this.  I have not pushed a
> release in a while, but I will rely on Gary's kind help to get this
> done assuming others are amenable.

This is a slippery slope and we need *clear* rules. Some users have
already asked Commons to release `commons-lang3` 3.17.1 [6] because
their policy doesn’t allow upgrading to a new *minor* version.

Piotr

[1] https://lists.apache.org/thread/t940gc5b5g7k176ry47okzhlv29y03k1
[2] https://github.com/apache/tooling-trusted-releases/issues/183
[3] https://github.com/apache/tooling-trusted-releases/issues/222
[4]
https://docs.google.com/document/d/1IZnHEwzz1N7LbChVkZTE_dfo3I2np8rULssq5I2wchM/
[5] https://github.com/vex-generation-toolset
[6] https://github.com/spring-projects/spring-boot/issues/46437


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to