Re: pkg_info -X: Add FILE_CKSUM (sha256)

Taylor R Campbell <[email protected]>
Newsgroups gmane.os.netbsd.devel.packages
Message-ID <[email protected]>
> Date: Sat, 17 Jan 2026 20:45:26 +0000
> From: Taylor R Campbell <[email protected]>
> 
> > Date: Sat, 17 Jan 2026 09:44:11 +0000 (UTC)
> > From: Benny Siegert <[email protected]>
> > 
> > On Sat, 17 Jan 2026, Jonathan Perkin wrote:
> > 
> > > Not yet.  Taylor has an open PR for this, but I owe him a review (I don't 
> > > like making it an error for pkg_summary not to have checksums).
> > 
> > Makes sense. The behavior I would like to see (I think) is:
> > 
> > If there is a checksum line in the summary, check the sum and refuse to 
> > install the package if it does not match.
> > 
> > If there is no checksum line for the package in the summary, skip 
> > checking.
> 
> Here is what I proposed for pkgin:
> 
> https://github.com/NetBSDfr/pkgin/pull/141
> 
> > Refuse to download packages without hashes, unless the environment
> > variable PKGIN_ALLOW_UNHASHED is set.

I thought about it further and decided to remove this part.  With the
updated change to pkgin, if the pkg_summary file has SHA-256 hashes,
they will be checked; if it doesn't, then they won't be, as bsiegert
and jperkin both suggested.

This way, there is no behaviour change anywhere, unless the package
builder uses pkg_info -H sha256 to generate the summary, at which
point -- if the change is merged -- newer pkgin will automatically
start verifying the hashes as it downloads package files.
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.