On 3/23/10, Mark Cave-Ayland mark.cave-ayland@siriusit.co.uk wrote:
Andreas Färber wrote:
my SPARC64 Milax ISO test now crashes [...]. However, [...] part of me
feels I should commit what I've got so that other people can get eyeballs on this and help out.
Please don't commit known-bad refactoring steps, that's really bad for
bisecting.
If you want more review, consider pushing a Git branch somewhere. Note
that most patches sent to this list are not git-am'able so need special care when applying, like manually undoing quoted-printable...
Thanks for the response Andreas.
I'm very happy that the patch is the way forward given the current diversity within the OpenBIOS codebase. I'm just nervous that a longer timeframe means that other patches will causing merge difficulties in the future for what is already quite a large patch (see attachment with my last email).
I think the best that I can do so far is to commit when it passes my tests which are primarily on SPARC64. In terms of other platforms, I've asked more than once how I can test them before commit but no-one has offered any responses. Then again, as a result of the new refactoring I would expect fixes to be a lot simpler.
I don't think I have any ELF test cases, unless Solaris happens to use our ELF loader during the second stage.
Maybe it would make more sense for OpenBIOS to move over to git for development testing - but then again, commits of this size will undoubtedly be quite rare. What does everyone else think?
I like git, though with git-svn I can work with my tree with StGIT, QGit and git tools as if the main repository were also git.