Adam Heath writes ("Re: Disputes between developers - draft guidelines"):
> I think our interactions are germane to the project, so, I'm adding a cc to
> the bug, as you should have done(geez, aren't you going to even follow your
> own guidelines?)I meant to, but I forgot. I've sent the BTS a separate copy now. > > Perhaps you missed my mails asking you where this previous discussion > > was ? I'm afraid I still can't seem to find it. > > Well, I can't find it either. I know it has been done, tho. If you can't provide me with a reference I'm afraid I'm going to have to ask you to repeat yourself. > > For example, I said in my mail of Wed, 16 Oct 2002 20:03:17 +0100: > > > > The fact that textutils md5sum has this bug is the reason we're not > > using it. (a) this is a gratuituous change and (b) it makes the > > program less useful. > > Er, it's not a bug in textutils. It never was. It's a bug in your program > for not handling textutil's output. By `my program' you mean the program that's calling md5sum ? I disagree. dpkg's md5sum comes from PGP2, and predates textutils' version by some time. If there is any kind of de facto specification, then the pgp/dpkg behaviour has to be definitive. > > (a) It is not backwards compatible. Existing software which uses > > md5sum and does something with the output will break (depending on the > > exact circumstances). > > So, we can never ever possibly minutely even try to change? I don't *think* > so. Well, if the change were a good one, then perhaps there would be some point in the pain of transition. But what purpose does the change serve ? > > (b) The new behaviour leaves no way to just get the unvarnished md5sum > > of some file. Being able to do this is quite useful if you're writing > > shell scripts and the like. > > How hard is cut -f 1 -d ' ' or sed -e 's/ .*//'? It's needless complexity. Why have md5sum go out of its way to do something that's actually unhelpful ? Ian.
