Re: squid master build with alternate openssl fails

On 5/8/20 5:11 PM, Francesco Chemolli wrote:
> I rebuild our docker
> images once a month to ensure they're fresh with what's in the wild.

FWIW, I think montly update frequency is excessive when there is no
adequate update validation. I speculate that we are spending more time
on faulty build tests than on fixing true build bugs the tests expose.


> To decide how much effort to invest, how prevalent is this situation? On
> Linux I'd expect this to be pretty much a corner case by now, is it not?

Unfortunately, we do not know. Folks running SslBump on older OS
releases may want to build with newer OpenSSL releases. Is that 1% of
Squid deployments? 10%? 20% of those deployments that matter? Unknown.

If it takes more than a few hours to change an existing (or add a new)
test node that uses custom library locations for --with-foo=PATH
options, then I would not do it (for now). You have bigger fish to catch.

Alex.
_______________________________________________
squid-dev mailing list
[email protected]
http://lists.squid-cache.org/listinfo/squid-dev