Bart Smaalders wrote: > Garrett D'Amore wrote: >> I have two questions: >> >> 1) Given that pretty much the entire FOSS universe seems to be moving >> away or have already moved away from, CVS, should we perhaps consider >> integration as Obsolete, with some suitable note in the man pages or >> other docs recommending the use of alternatives such as SVN or Hg? >> > > The entire FOSS universe?
See "pretty much", not "all". > > http://xml.apache.org/cvs.html Apache uses SVN for most of its projects. Only a few of them seem to be continuing to use CVS. > http://developer.apple.com/opensource/cvs.html I wonder if Apple uses CVS internally. The developer.apple.com website seems much more geared towards use of CVS to "push" software, than as a true SCM. > http://cvsweb.netbsd.org/bsdweb.cgi/ NetBSD has had numerous projects in process attempting to replace CVS. I would definitely count them as "moving away". Part of the problem is that they haven't decided *which* SCM to settle on -- or at least they hadn't as of about a year ago. > > I still find older project using this; not having it available is an > impediment. > > Changing the docs to suggest alternatives is not really part of a man > page. It might be a reasonable comment in a developer doc entitled > "choosing a SCMS"; such a document doesn't exist as far as I know. I beg to differ. There are numerous examples where man pages indicate that a tool, API, or otherwise is obsolete and "suggests" possible alternatives. Folks who need CVS for existing projects will have it available for their use. I agree that this is good. But at the same time, I don't think it is a terrible idea to gently steer folks towards other SCMs, particularly for new projects, where we (OpenSolaris and Sun) may be better equipped to provide support. (Better equipped because we use the SCM ourselves, because the SCM doesn't suffer from one of the known defects of CVS, and because the SCM also has folks still actively maintaining it.) > >> 2) Not really architectural, but you indicate that this is released >> under GPLv1. If so, are there are any other GPLv1 software in >> OpenSolaris? Maybe this needs a special case of legal review? (I am >> not entirely familiar with the differences between GPLv1 and GPLv2....) >> > > I would assume the OSR process will explore any issues here. I suggest > that if you're worried about this, _you_ do the background reading > and have the discussion w/ legal, rather than raising vague "fetch > a rock" questions such as asking the project team to explore if any > other components in OpenSolaris that use this license. Actually the project team has already answered this to my satisfaction. The main point here was concern that precedent set for other GPLv2 software *might* not apply for GPLv1. But an OSR has already been done, so I'm quite happy on this point. -- Garrett
