Brock Pytlik wrote: > John Rice wrote: >> Great that looks simpler again - we'll make the changes when are back >> in on Tuesday, test to make sure this does what's needed for both the >> GUI clients and respin. If Danek hasn't found a fix for the >> __init__.py issues we'll revert this file to r590. >> >> Bug logged to track root cause analysis of problem post 2008.11 (4116 >> depends on this): >> 4231:Resolve root cause for memory leak and remove workaround for 4116 >> >> Brock you happy with this approach, as opposed to modifying the API? >> > Actually, I find modifying the api cleaner than this. This just makes my > skin crawl personally. I'm not inherently opposed to leaving the API > workaround in for a while we get things fixed up. I think an API option > of "prefer low memory use over blazing speed" is at least a possibly > reasonable switch to provide. Hard coding in our pkg client name here > seems wrong to me on many levels. If this is what everyone else likes, > I'm fine with it for 2008.11, but this workaround I want out of the code > as soon as possible.
The client_name is already hard-coded, and as I stated, this is only intended as a temporary change. I agree that it might be nice to have the api have a minimal-memory-usage flag. But until we can adequately design that, I'd rather not change the api to do it. -- Shawn Walker _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
