Jyri Virkki wrote: > Gopalakrishnan Sankaran wrote: > >> Here is the webrev after making the changes required (as suggested in >> the initial round of review) >> Kindly review the changes. >> >> http://cr.opensolaris.org/~gs193864/cr6644643/ >> > > Hi, > > Please let us know which set of gems did you use for testing the > native compilation? >
I'm not speaking for this code review in particular(and I asked Gopal this same question in another email which he needs to answer). We tested the rbconfig.rb POC with MySQL, Postgres, EventMachine and Fastthread - which are some of the popular gems. we also did some performance testing to quantify the increased overhead(which was negligible, when present). -ps > > - Synopsis of CR 6644643 is > "rbconfig.rb contains bogus paths, breaks gems requiring C compilation" > but the diffs show > "bad paths in rbconfig.rb file" > Make them consistent (pick one or the other). > > > *** usr/src/cmd/ruby18/Makefile.sfw > - Something is wrong in your workspace, as the diff doesn't show a change > in the version (ident string). Please find/fix. > - Use $(GMAKE) instead of hardcoding the path. > > *** usr/src/cmd/ruby18/postinstall-ruby > - Something is wrong in your workspace, as the diff doesn't show a change > in the version (ident string). Please find/fix. > - This introduces a dependency on gcc for the ruby package, so you should > update the package 'depend' file accordingly. > > *** usr/src/pkgdefs/SUNWruby18u/prototype_* > - Something is wrong in your workspace, as the diff doesn't show a change > in the version (ident string). Please find/fix. > - How is a user who wants to use Studio cc expected to switch to the other > configuration? What happens next time they upgrade SUNWruby18u package? > The symlink isn't marked editable (AFAIK there's no support for that). > Have you tested this to make sure it works? > > > -- Prashant Srinivasan F/OSS Enthusiast Sun Microsystems, Inc. http://blogs.sun.com/prashant GnuPG key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x82FBDE5A
