Johannes Schindelin <Johannes.Schindelin@xxxxxx> writes: > If you feel strongly about your contrived examples possibly being > affected by this patch, we could easily make this conditional on > > 1) no '&&' or '||' being found on the command-line, and > 2) argv[0] not containing an '=' > > Another approach would be to verify that argv[i] starts with '-' for > non-zero i. > > But do we really need to do that? No. An explicit way to override an incorrect guess is sufficient and necessary. The above two-item list you gave will be just part of the machinery to make an incorrect guess. The auto-detection in the posted patch should cover many users' use case and I do not think it needs to be extended further to make it more brittle, as by definition its guess cannot be perfect. Just keep it simple and give a separate escape hatch. > That means that the user has to specify something like > > HAHAHA_IT_IS_NOT=/plink.exe ssh > > as GIT_SSH_COMMAND. My second message was to clarify that "VAR1=VAL2 command" is NOT a contrived example, and this response indicates that I somehow failed to convey that to you. The "if tortoiseplink exists (and the end user can override the location with an environment), use that, and if PuTTY plink exists (ditto), use that instead" in a "myssh" script, and use it as core.sshcommand with the environment to override my custom installation location to these two programs, would be what I would do when I get two Windows machines, with these variants of SSH on each. So take the second message as a bug report against the version of Git for Windows you ship with the patch in question. The auto-detection may work for many people and that is a great thing. I failed to say that in my message, as I thought that was obvious. But it is important to plan to cope with the case where it does not work. The usual practice around here is to say "the it may not necessarily work for everybody, so lets be prepared to add an explicit override if it turns out to be necessary". The second message, which you are responding to, was meant to be a bug report from the future, telling us that an override is needed, showing that we do not have to wait for a bug report to act on.