Hi Peff, On Wed, 31 May 2017, Jeff King wrote: > On Wed, May 31, 2017 at 05:27:21PM +0200, Johannes Schindelin wrote: > > > > My intent in putting it into the actual git binary was that it could > > > also be useful for collecting build-time knobs from users (who may be > > > using a binary package). Like: > > > > > > http://public-inbox.org/git/20160712035719.GA30281@xxxxxxxxxxxxxxxxxxxxx/ > > > > > > We haven't filled in that NEEDSWORK yet, but I'd rather see us go in > > > that direction than remove the option entirely. > > > > FWIW it also helped Git for Windows. > > > > The two additional bits we added to the build options (the commit from > > which Git was built and the architecture) did not hurt one bit, either. > > > > In other words, it would make my life a lot harder if --build-options were > > moved to a test helper that does not ship with the end product. > > Cool, I'm glad it has helped already. If you have further bits added to > the output, is it worth sending that patch upstream? Yes, of course. The day only has 24h though and I am still stuck with other things I try to contribute that seem to be requiring a lot more effort (mostly trying to make my case that there are certain coding paradigms I find too sloppy to put my name on) from my side to get accepted than I'd like. So yeah, as soon as the queue drains a bit more, I have tons more patches ready to go upstream. Ciao, Dscho