On Tue, Oct 30, 2007 at 11:55:20PM +0100, Eric Kow wrote: > Some miscellaneous results of our issue553 debugging with David. My > implementation of David's Workaround-enforcement idea is to make it so that > you > have to import the functions qualified. This is a little bit icky, but it > does > have the extra benefit of making us extra-conscious where our code comes from. > > I'd send the rest of it (actually using the enforcement) after my issue298 > stuff goes in, as it will depend on it. No point adding junk to the bundle. > > Tue Oct 30 21:42:56 CET 2007 Eric Kow <[EMAIL PROTECTED]> > * Simplify writeHashFile function. > > Tue Oct 30 23:28:55 CET 2007 Eric Kow <[EMAIL PROTECTED]> > * Modernise and enhance buggy renameFile workaround. > > Use the hierarchical library structure. > Catch only 'does not exist' errors, passing everything else up. > > Tue Oct 30 23:49:59 CET 2007 Eric Kow <[EMAIL PROTECTED]> > * Add ability to enforce use of Workaround functions.
This all looks good. I haven't reviewed the last bit of code for correctness, because it's a bit daunting. I wonder if we should switch to some sort of ghc-api or haskell parser for the policy checking? I don't relish the idea of searching for import statements in shell. Note also that we should disallow blanket imports of dangerous modules (e.g. import System.Directory should be an error). -- David Roundy Department of Physics Oregon State University _______________________________________________ darcs-devel mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-devel
