On Fri, Sep 30, 2011 at 01:29, Jim Warner <james.warner@xxxxxxxxxxx> wrote: > On Sep 29, 2011, at 6:04 PM, Craig Small wrote: > > I have no clue as to why or how the original merge history can > be made permanent. Perhaps Sami knows. I don't know how real maintainers do the thing, but the list bellow 'sounds' correct to me. I get the impression that maintainers 1. have a private remote repository where own coding efforts are kept 2. the own repository is the one which you push by default 3. other repositories are fetched, including people who are active contributors 4. upstream is a repository is used explicitly 5. committing to upstream/master is ok, if patches are coming in `git format-patch' format, or one genuinely writes change to master branch (applying diff is fine, but not recommendable because author & date tends to get messed) 6. there is master branch in upstream, and not many other branches 7. upstream does not change history; no forced pushes, no rebases, no tag changes, preferably no branch deletions... http://changelog.complete.org/archives/586-rebase-considered-harmful Warning; that is what I think the real maintainers do, I have never been one so my advice is not more than me thinking aloud. Having some seasoned maintainer & git guru telling 'yup, that is approximately right, and you should consider also [...]' would be splendid. >> The oom flag should be a configure option. Either --enable-blah >> or --disable-blah depending if it is default on or off. > > It'll be in ./configure --help after you run autogen.sh. Look > for '--enable_oomem'. It is now --enable-oomem. The oomem field could be generally preferred so in future one might want to change that to --disable, but let's get the merge done first. -- Sami Kerola http://www.iki.fi/kerolasa/