On Thu, Mar 05, 2009 at 14:22:18 -0500, Mark Stosberg wrote: > > What are the arguments for keeping libwww support (as opposed to > > focusing on libcurl and HTTP)? > > Outsourcing the http work to libcurl allows the darcs project to focus on it's > core compotencies: distributed source control. It seems like a good course > to outsource with reasources are limited.
Indeed. Right now, we outsource to one of - libcurl - libwww - HTTP (Haskell) - wget/curl command line (fallback) So what I was asking the list is if there was a particular reason to keep libwww support. Libcurl support is understandable; it's stable and featureful. HTTP support is somewhat understandable; it's native Haskell and seems to be easier for us to use (I'm just saying that because some functionality like maintainer file checking just seems to depend on the HTTP package outright, which I guessing is because it was simpler to do so) The motivation behind libwww support is less clear. I suspect it may have been something to do with pipelining (which I understand libcurl supports). I'm just hoping we can further this trend of whittling away at our code, removing needless choice and variation along the way... -- Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow> PGP Key ID: 08AC04F9
signature.asc
Description: Digital signature
_______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
