On 19.09.2019 17:51, Robert Elz wrote: > I think the actual change is fine ... just 2 caveats. > > First, the comment in the patch will be meaningless after the patch > is applied - that comment belongs being the cvs commit message for > the log, not part of the resulting file (which could have a comment > explaining what is happening - but not one that says why the old way, > which no longer exists, isn't correct). > I will skip the comment in the code. > Second - please attempt to batch up changes to sys/cdefs.h ... this one > is not urgent. You should fix it locally, use the fixed version to > find uses which are incorrect, and fix (change to some other assert, or > simply remove) the broken ones, and commit those fixes - but wait until > there is some other need to commit the change to cdefs.h (or when someone > else has just committed a change there) - every time that file changes > everyone's update builds tunr into (effectively) full builds, as just > about every .c file we have depends upon cdefs. (it is even worse than > changes to sys/param.h from this POV). > I'm waiting for riastradh@ to fix the reports in DRMKMS and nist_hash_drbg.c. There are no other ones (at least in my build configuration). > kre >
Attachment:
signature.asc
Description: OpenPGP digital signature