Hello Karsten, thanks for your analysis. Most of the patches you refer to are simply switching off tests for MINGW; let me comment on the remaining ones: > * t0110-urlmatch-normalization: 1 > > Passing binary data on the command line...would have to > teach test-urlmatch-normalization.c to read from stdin or file. > https://github.com/msysgit/git/commit/be0d6dee Indeed, that would be better solution. For now, I'm going to submit the switch-off patch you mention. > * t4036-format-patch-signer-mime: 1 > > Passing non-ASCII by environment variable, will be fixed by Unicode > environment support. Will submit that patch series soon. > * t7001-mv: 6 > cp -P fails in MinGW - perhaps use the long option forms (--no-dereference)? "cp -P" fails with our 2001-edition of cp, so msysgit had to revert: https://github.com/msysgit/git/commit/6d3e23d4 But I was ashamed to mention that upstream; and I hope mingwGitDevEnv is going to solve that. > * t8001-annotate/t8002-blame: 5 > > Msys.dll thinks '-L/regex/' is an absolute path and expands to '-LC:/msysgit/regex/'. > https://github.com/msysgit/git/commit/2d52168a Nice! But I'm afraid the patch cannot be submitted upstream as it is. I think the hack could be automated by processing options "-L*" this way: sed 'sX\(^-L\|,\)\^\?/X&\\;*Xg' Then it would become only few lines at the top of the script, executed on mingw only. I hope to submit the patch in this form soon. Have a nice day, Stepan -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html