Re: RFS: coffeescript

2010-04-02 Thread Geza Kovacs
On 04/02/2010 05:42 AM, Christoph Egger wrote:
 I've added descriptions to the patches, as recommended by DEP-3. I'm not
 forwarding the patches upstream, as they mostly relate to
 debian-specific quirks and are not generally useful (specifically, one
 deals with Node.js being installed to /usr/bin/nodejs in Debian instead
 of upstream's default location at /usr/bin/node, and the other is simply
 to allow coffee to be installed directly in /usr/bin/coffee instead of
 /usr/lib/coffeescript/coffee, whereas upstream instead uses an
 unnecessary symlink).
 
   OK I suspected that. That's one of the important things that could
 live in a patch description/header ;)
 

I've added the Forwarded: not-needed tag (described in DEP-3) to the
patches to indicate that they aren't meant to be forwarded upstream.

   debian/rules is just a makefile, you can call chmod from there.
 

I have added an override for dh_fixperms that does this.

Thanks,
Geza


-- 
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4bb6821a.2090...@mit.edu



Re: RFS: coffeescript

2010-04-01 Thread Geza Kovacs
On 04/01/2010 06:03 AM, Christoph Egger wrote:
 Quoting Upstream Website:
 
 Disclaimer: CoffeeScript is just for fun. Until it reaches 1.0,
 there are no guarantees that the syntax won't change between
 versions. That said, it compiles into clean JavaScript (the good
 parts) that can use existing JavaScript libraries seamlessly, and
 passes through JSLint without warnings. The compiled output is quite
 readable — pretty-printed, with comments preserved intact.
 
   Do you think it's sensible to have that packaged in debian (and
 it's stable releases)? Looks rather like a volatile thing where one
 wants for a 1.0 before packaging to me (though that impressin is based
 on a quick look on the webpage so you may know better ;))
 

I'd agree with your worries if this were a purely interpreted language
like Python, where changes in the interpreter require entire program
rewrites so that existing deployments can still run. However, this is a
compiled language, which outputs to standard Javascript. Hence, even in
the unlikely case that a huge change makes existing programs
uncompilable (I say unlikely because coffeescript has done a good job of
maintaining backwards compatibility with releases), then the existing
compiled Javascript that has been deployed will still work. Since this
is primarily used for short-term web development, and not huge
multi-year projects, then the question of whether a program will still
compile years later is less of an issue.

Also note that Debian is already ripe with languages which haven't yet
reached 1.0 and are still adding new syntactic features, like boo. I
believe so long as there aren't Debian packages that depend on
coffeescript to get compiled, the availability of this package shouldn't
cause additional maintainability issues, and as with any cutting-edge
development tools, it should be left up to the user whether to use them
or not.


-- 
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4bb4aa73.7020...@mit.edu



Re: RFS: coffeescript

2010-04-01 Thread Geza Kovacs
On 04/01/2010 12:42 PM, Christoph Egger wrote:
  * Could you maybe merge the changelog entries into a single one?
There's no reason to increase the debian revision for every
bullet-point ;)

Done, I've merged the changelogs and am now back at 0.5.6-1

  * You're using format '3.0 (quilt)' for your package so applying the
patches works at extraction time and there's no need for a '--with
quilt' in the rules file (also, for a --with quilt you'd need to
build-depend on a correct version of the quilt package)

Done, I've removed the --with-quilt

  * Your patches don't have any description on them. I've missed such
information quite often when adopting some package (there's some
DEP for a uniform format somewhere). Also have you forwarded the
patches upstream?

I've added descriptions to the patches, as recommended by DEP-3. I'm not
forwarding the patches upstream, as they mostly relate to
debian-specific quirks and are not generally useful (specifically, one
deals with Node.js being installed to /usr/bin/nodejs in Debian instead
of upstream's default location at /usr/bin/node, and the other is simply
to allow coffee to be installed directly in /usr/bin/coffee instead of
/usr/lib/coffeescript/coffee, whereas upstream instead uses an
unnecessary symlink).

 
   Have you tested your package with lintian? I'm certainly not
 nit-picking on Information or Pedantic tags but some of the
 Error/Warnings definitely look worth fixing (invoking linitan with -i
 additionally gives a description of the issues at hand):
 

I've fixed as many of the warnings as I could; currently the output of
lintian is:

P: coffeescript-doc: no-upstream-changelog
I: coffeescript-doc: extended-description-is-probably-too-short
P: coffeescript: no-upstream-changelog
W: coffeescript: binary-without-manpage usr/bin/cake
W: coffeescript: unusual-interpreter ./usr/bin/cake #!nodejs
W: coffeescript: unusual-interpreter ./usr/bin/coffee #!nodejs
W: coffeescript: executable-not-elf-or-script
./usr/lib/coffeescript/optparse.js
W: coffeescript: executable-not-elf-or-script ./usr/lib/coffeescript/cake.js
W: coffeescript: executable-not-elf-or-script
./usr/lib/coffeescript/parser.js

The unusual-interpreter errors are of course unfixable since the program
only runs on nodejs, there is no upstream changelog file (though there
is one on the website, I could copy-paste it if desired but I don't
think that's in line with Debian policy), and the warnings about
executable permissions I unfortunately didn't figure out how to fix
since quilt doesn't seem to be able to keep track of file permissions
(if anyone has suggestions on how to store changes to the file
permissions in the quilt patch set do let me know).

Thanks,
Geza


-- 
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4bb55cfa.5020...@mit.edu



RFS: coffeescript

2010-03-31 Thread Geza Kovacs
(Please CC any replies to me)

Dear mentors,

I am looking for a sponsor for my package coffeescript.

* Package name: coffeescript
  Version : 0.5.6-6
  Upstream Author : Jeremy Ashkenas jer...@documentcloud.org
* URL : http://coffeescript.org
* License : MIT
  Section : devel

It builds these binary packages:
coffeescript - interpreter and compiler for the CoffeeScript language
coffeescript-doc - documentation for coffeescript

The package appears to be lintian clean.

The package can be found on mentors.debian.net:
- URL: http://mentors.debian.net/debian/pool/main/c/coffeescript
- Source repository: deb-src http://mentors.debian.net/debian unstable
main contrib non-free
- dget
http://mentors.debian.net/debian/pool/main/c/coffeescript/coffeescript_0.5.6-6.dsc

I would be glad if someone uploaded this package for me, or provided
feedback on any issues with the packaging so I can fix them.

Kind regards
 Geza Kovacs

(Please CC any replies to me)


-- 
To UNSUBSCRIBE, email to debian-mentors-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/4bb40884.80...@mit.edu