On 10/30/2012 04:48 PM, John Dennis wrote:
On 10/30/2012 10:41 AM, Petr Vobornik wrote:
I would rather pack it with FreeIPA. I would bundle the Dojo builder
either way because is not part of the Fedora package. Other thing is
that the build in the fedora package won't be probably suited for
FreeIPA - it's not build into single file.
Ugh. Do we really want to be in the business of packaging and supporting
a complex framework exclusively for IPA?
I wan't to bundle only custom, minimal build of the framework (some
parts of dojo/dojo) and a builder (just in source rpm). To not put us
into situation that we would have to support whole dojo in all possible
By using system package would have to support whole lib in all systems
we build for.
slightly different case. They are transmitted over net to client's
browser and therefore the amount of files and their size have to be kept
I know there has been a move afoot for "software collections" where
packages install with their own private copy of some other package
because they can then control the dependency. But that opens up a whole
raft of problems (security patches to name just one) that I think we
would be better off avoiding. IMHO system supplied packages are a clear
winner over private copies of packages but I recognize not everyone
agrees with this.
There was a proposal(or just idea) for Fedora to change a JS packaging
* js libraries will be packed in separate rpms
* package with Web App would declare it's dependencies
* when the package would be installed or dependency updated, it would
copy the dependency's files and incorporate (build) them according to
What is the chance other projects in the distros we work with make the
decision to use Dojo? Do any of them already do? What is the experience?
I don't know about any. IMHO we can only guess. There are big
differences in developers preferences. Some rather use more server
technology like some jboss-based-framework or django, some a single page
application (our web UI).
Just as an aside. I'm not a web developer but from standing on the
sidelines it seems to me every 6 months there is a new web framework
djour. It makes my head spin, this is the most volatile area of software
I've ever seen. It seems like as soon as you've hitched your horse to
one you're behind the curve because of the next best thing. I wonder why
this software domain can't settle on one or two basic technologies
instead of the constant churn and disruption. Wouldn't that benefit
Web undertook a major shift in last, lets say, 5 years. Many
technologies appeared and many will be forgotten. The questing is which
ones to choose. One of the arguments for Dojo in this area is that it
isn't a 'hip' thing. It has been here since 1997 and it's still evolving
based on current needs.
Freeipa-devel mailing list