Antoine just gave me the results for an exp-run with this change applied: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=231256

In D17040#364634, @rgrimes wrote:

@ed can we get this added to the CI work load as a once a week type run? IIRC it takes a 16 core machine about 3 days to run build_option_survey.sh

Hi Conrad,

It looks like I'm quite the forgetful person. In D9169, cem@ tried to solve the same thing. Eventually he abandoned the idea, because we weren't able to accurately implement the __generic() matching to filter out bad calls.

Add logic to explictly make old calls to qsort_r(3) fail.

It's a fair point you two are making: with the change I sent out, code that calls qsort_r() in the existing way will compile without any warnings/errors. This is why I've just changed my patch to add some logic to stdlib.h and libkern.h to detect bad calls, similar to how we dealt with the dirname() / basename() transition (libgen.h). Calls to qsort_r() with the comparator as the last argument will fail explictly.

Uriparser 0.8.6 has been released in the meantime.

Looks good. Thanks a lot for working on this!

Ben: any final thoughts before I commit this?

Process woodsb02's feedback.

Hi Ben,

Process feedback from kib@.

Don't conditionalize the addition of the pointer.

Always show the <PRI>VERSION. Don't let it depend on LogFacPri.

Oops! Accidentally used the wrong differential revision in the commit message. This was committed as r332100.

Hi Ben,

Make the scope of what is affected a bit more explicit.

Thanks for your input, Kostik. What do you think about this version?

Extend the UPDATING note.

Add a note to UPDATING regarding this change.

In D14951#314438, @ed wrote:

There are some alternatives, such as syslog-ng and rsyslog. Documentation indicates that these also support RFC 5424 properly. I will, however, check this to be sure.

Process feedback provided by kib@.

Hi Kostik,

Make minor changes based on woodsb02's feedback.

Respect RemoteAddDate (-T) for RFC 5424 messages.

Process the excellent feedback given by woodsb02@. Thanks!

Add explicit buffer size to parsemsg_remove_unsafe_characters().

Looks good to me. Thanks!

While there, make use of GitHub instead of SourceForge.

Revert unnecessary changs to other ports.

Is there a need to describe how these should be formatted? Can we place multiple initializers on one line, or should they all be placed on a separate one? Tabs before the = to align them?