node: pre-freeze upload?

[Carsten Hey <carsten@debian.org>]

Hi,

I prepared a possible node "team upload", i.e., an upload by someone not
listed in the uploaders field with permission from the maintainers.

It updates the packaging from debhelper compat level 2 to 8, updates the
build-system to ease switching command names (including the name and
content of the man pages) and eases renaming the package.  A possible
renaming after the freeze would, additional to the changes required for
a sane migration, only require these three names to be changed (one in
debian/control and two in the options used to configure node).  In
related news, metapackages would the proper section for a symlink
providing package to prevent autoremoving $newpackagename, even though
it is not a meta package.

I rebuilt node 0.3.2-7.1 and compared the resulting .deb with the .deb
from the prepared 0.3.2-8 (I used the same directory name for building
both packages! I don't use exclamation marks very often.).  In
control.tar.gz, the files conffiles (different order), control and
md5sums are different.  In data.tar.gz, only the copyright (lintian
fixes) and the changelog differ.

Due to the changes in upstream's build system (including renaming some
files), the interdiff in the two .diff.gz is huge and hardly anything
the release team would be happy to review and an interdiff it presumably
would like is nothing that could be forwarded to upstream.  Everything
in the .diff.gz except of etc/* and debian/* belongs to the changes in
the build system.

I am not able to test node, but if rebuilding 0.3.2-7.1 results in
a working package, then the prepared 0.3.2-8 is fine too.  It would by
great it you could upload 0.3.2-8 based on [1] or if you could test the
package and tell me to upload it.

Things not done/applied:
 * The package links to zlib without a reason, this could be fixed.
 * The patch [2] would unclutter the compiler output, but change the
   binary's Build-Id the way Debian packages are normally built.
 * The patch [3] is used in Fedora since ages.  It could be checked and
   possibly applied.
 * Lintian proposes some minor fixes in the man pages.


Regards
Carsten


 [1] http://people.debian.org/~carsten/node/node_0.3.2-8.dsc
     Checksums-Sha256:
     adbe7098815470076906e7a82c406e792ac2c87015497dd7172b23d38c2e4507 900 node_0.3.2-8.dsc
     5707846c2d9de3c0feee75987a58739cae0bdcc97176915e3953b2cfb4a0a7f3 14693 node_0.3.2-8.diff.gz

 [2] http://people.debian.org/~carsten/node/rename_log_to_log_msg.patch

 [3] http://people.debian.org/~carsten/node/node-0.3.2-rose.patch