Re: More - vs . woes
Hi Charles,
On Sun, Jun 10, 2018 at 08:37:29PM +0900, Charles Plessy wrote:
> so if I understand well, when a R upstream author choses to release his
> package with a given version string (given by the Version field in the
> pachage's DESCRIPTION file; let's call it the "canonical" version
> string),
I like the term "canonical" version string.
> As I have not yet read a good reason for using that feature, I think
> that it is definitely worth asking upstream authors whether they would
> consider using only "canonical" version strings in their DESCRIPTION
> files.
>From my perspective is the diversion of a given string a typo. There
are lots of typos that do not harm and will go unnoticed. I consider it
good style to point upstream to some typo anyway, specifically if there
are cases where it really does harm. If upstream would respond by "no,
that's no typo but intentional since it has advantage xy" I could
perfectly work around. Since in all my practice upstream was happy
about pointing out typos and has fixed it (if there was an active
upstream) I'll keep on pointing out upstream kindly (and I'm hoping for
your input to work on the "kindly" feature).
> Thanks for the emails you sent. (Perhaps a little bit more
> background about Debian in the email might be useful.)
I'd welcome actual enhancements / patches to my mails about this.
> If it is not just isolated typos, or if the problem is causing us
> troubles regularly, one possible solution would be "normalise" all R
> version numbers in Debian packages, for instance by always replacing
> dashes.
I agree with Dirk in this specific point that its not a good idea to
change the upstream version string. IMHO it has brought more harm than
good when we did so.
Kind regards
Andreas.
--
http://fam-tille.de
Reply to: