[Pkg-javascript-devel] libjs-* vs. libnode-*

[Jonas Smedegaard]

On 11-10-14 at 12:02am, Jérémy Lal wrote:
> On 13/10/2011 23:48, David Paleino wrote:
> > On Thu, 13 Oct 2011 20:45:34 +0200, Jonas Smedegaard wrote:
> > 
> >> Hi,
> >>
> >> I just noticed that libnode-backbone-dirty depends on a mixture of 
> >> libjs-* and libnode-* packages.  That is correct according to 
> >> Debian Policy, but looks ugly.
> >>
> >> We discussed it briefly recently, where Jérémy felt that it was 
> >> unusual for same package to support both libjs-* and libnode-* so 
> >> less of an issue.
> >>
> >> Now that David has added a bunch additional packages there is 
> >> perhaps material for looking at this issue again?
> > 
> > Given I hadn't found any policy about this, I went "by heart" about 
> > package naming.
> > I called libnode-* those libraries I thought were node-specific, and 
> > libjs-* those which could be of general usage (example: 
> > libjs-backbone).
> > 
> > Still, I'm not satisfied by the approach I'm taking, and would 
> > prefer a more objective guideline :)
> 
> I think libjs-* should be reserved for "browser libs" uses.

> In some cases, the choice is obvious, since the lib cannot run, or has 
> no use, inside a browser. Then it should be named libnode-*.
> There are cases where it can run on both : i'm for building two 
> binaries out of one source.
> Symlinks to the libjs-* from libnode-* may not be a good idea. It 
> seems simpler to just build two independent packages.

I agree that it generally makes sense to package separately:

For browsers it makes sense to aggresively merge and compress sources 
using uglifyjs (or yui-compressor), as bandwidth is a major factor and 
compressors test for _browser_ compatibility.

For Node it makes sense to install code snippets unmerged and 
uncompressed, allowing Node to deal with it directly - perhaps improving 
performance by tracking timestamps or location of each snippets.

Might be that some package really is identical for both browser and Node 
consumption - but that is most likely the exception, not the rule.

I do recall now that something along that reasoning was raised by Jérémy 
in the past - and that I didn't understand it.  Now I do.  So credits to 
Jérémy for above!


Someone volunteer to create a wiki page to document this?

I suggest that we call it "Debian Javascript Policy" because browser and 
Node code are both subsets of Javascript.  But again I recall Jérémy 
arguing differently in the past.  Care to try elaborate again, Jérémy?


 - Jonas

-- 
 * Jonas Smedegaard - idealist & Internet-arkitekt
 * Tlf.: +45 40843136  Website: http://dr.jones.dk/

 [x] quote me freely  [ ] ask before reusing  [ ] keep private
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-javascript-devel/attachments/20111014/3e5eb0ee/attachment.pgp>