The checksum is not a tampering countermeasure. It is a "mirror ran out of diskpace" or "IP checksums are only 32 bits" countermeasure.
On Thu, Aug 31, 2017 at 11:35 AM, Julian Hyde <jh...@apache.org> wrote: > As security experts, you and I know that. But Joe User maybe only checks > one digest. > > (Aren’t we all Joe User sometimes?) > > Julian > > > On Aug 31, 2017, at 11:30 AM, Mike Jumper <mike.jum...@guac-dev.org> > wrote: > > > > On Aug 31, 2017 11:21, "Julian Hyde" <jh...@apache.org> wrote: > > > > After downloading artifacts, there are 3 things to check: (1) the > download > > is successful; (2) the artifacts were indeed created by the named author; > > and (3) the artifacts have not been tampered with. > > > > A security expert would know to use the .md5 for (1), the .asc for (2), > and > > the .sha256 or .sha512 for (3). > > > > > > If there is a danger that the artifacts may be tampered with, there is an > > equivalent danger that the checksum files will be tampered with, as well. > > Checksums alone cannot be relied upon to verify an artifact hasn't been > > altered. > > > > Only the signature allows verification of authorship and integrity ... > > assuming users have secure access to the corresponding public keys, and > > that those keys are linked into the web of trust. > > > > - Mike > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org > For additional commands, e-mail: general-h...@incubator.apache.org > >
