Re: Bug#164889: md5sum <FILE produces spurious ` -' in output
>>"Raul" == Raul Miller <moth@debian.org> writes:
Raul> On Mon, Nov 11, 2002 at 03:31:51PM +1000, Anthony Towns wrote:
>> debootstrap broke. Things of the form:
>> x=`cat foo.changes | grep blah.*deb | cut -d\ -f2`
>> y=`md5sum < blah*.deb`
>> if [ "$x" != "$y" ]; then
Raul> Hmm...
Raul> There are essentially two approaches to this class of problem.
Raul> [1] require that md5sum not break this kind of code.
Raul> [2] require that this code be modified to work with the "new" md5sum.
Then we need a transition plan.
Raul> If we're going to go with [2], we can require a change of the form:
Raul> y=`md5sum < blah*.deb | awk '{print $1}'`
Fails for md5sum /some/file >> hash-file-to-be-parsed-later
Raul> Since making this kind of change doesn't break use of the
Raul> existing md5sum, I think we should recommend that all apps
Raul> using md5sum test using the gnu variant and be modified so they
Raul> work with either version. In particular, we should allow bugs
Raul> to be filed against packages which don't support the gnu
Raul> version.
wishlist bugs?
Raul> Until all those bugs have been closed -- and until every package with
Raul> a md5sum dependency has had a chance to be tested -- we should forbid
Raul> deployment of the gnu version.
How do we keep track of these bugs?
Raul> Also, taking a step back:
Raul> Once we're done with this specific issue, maybe we should think about
Raul> making this kind of phase-in process an option for package developers
Raul> to introduce on their own. In other words, allow people to say:
Raul> a I'm going to change XXX in a way that breaks the existing interface.
Raul> b You have a package with a dependency on my packages which
Raul> contains XXX.
Raul> c Here's one way to write code that will work with either interface.
Raul> d Here's where to get an instance of the new interface for you to test
Raul> against.
Raul> e [stuff about bug filing/closing and the eventual release of the new
interface>
Raul> Comments? Opinions?
This requires way more thought than this. Proper dependency
handling isa complicated; and a generic, scalable solution takes eons
to hammer out.
Raul> One thing that bothers me about these kind of migrations is that older
Raul> packages aren't going to know about interface changes introduced in
Raul> newer packages. This almost requires a [potentially large and ugly]
Raul> versioned "Conflicts:" in the package which supplies the new interface.
Also: We need to worry about upgrades. We used to worry about
partial upgrades -- people running mostly stable but upgrading a few
packages from testing.
These conficts against versions of packages that depend on on
one could get impossible to collate; unless tools are provided that
can walk all the dependency relationships and provide one with a
current list (keeping in mind that some of these packages may also
provde a versioned dependency on the package whose interface is
changing).
I think this is going to be impractical unless for the very
seldom used packages.
Ideally, the packages should provide a virtual package with an
interface version embedded in them (md-interface-1.1), and packages
depend on that interface; but even that fails to scale for packages
that may provide multiple interfaces, only some of which are
changing.
manoj
--
Graduating seniors, parents and friends... Let me begin by reassuring
you that my remarks today will stand up to the most stringent
requirements of the new appropriateness. The intra-college
sensitivity advisory committee has vetted the text of even trace
amounts of subconscious racism, sexism and classism. Moreover, a
faculty panel of deconstructionists have reconfigured the rhetorical
components within a post-structuralist framework, so as to expunge
any offensive elements of western rationalism and linear
logic. Finally, all references flowing from a white, male,
eurocentric perspective have been eliminated, as have any other
ruminations deemed denigrating to the political consensus of the
moment. Thank you and good luck. Doonesbury, the University
Chancellor's graduation speech.
Manoj Srivastava <srivasta@debian.org> <http://www.debian.org/%7Esrivasta/>
1024R/C7261095 print CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
1024D/BF24424C print 4966 F272 D093 B493 410B 924B 21BA DABB BF24 424C
Reply to: