Re: [PATCH 2/4] compat: use git_mkdtemp()

From: Jeff King <peff@peff.net>
Date: Dec 6, 2025, 02:11 UTC
Message-ID: <20251206021122.GC1714099@coredump.intra.peff.net>
In-Reply-To: <aebd0ffe-7914-4731-8f79-830bd3b5a147@web.de>

On Fri, Dec 05, 2025 at 01:11:40PM +0100, René Scharfe wrote:

Show 11 quoted lines

> > This one is a conditionally-compiled wrapper for NO_MKDTEMP. But since
> > we always have git_mkdtemp() available (as of your first patch), can't
> > we just point at it directly with the macro?
> 
> A worthwhile cleanup if we stop at this point, but complicated by
> targeting three build systems, the CMake build being broken on macOS and
> me only knowing how to fake NO_MKDEMP for make, leaving half the build
> space untestable for me.
> [...]
> At the very least this cleanup should be done in a separated patch, as
> it's harder than it looks.

OK, I am convinced that it is not entirely trivial and can go in a separate patch. :) Mostly I was surprised that you would not have followed through on an obvious cleanup opportunity. It just turned out harder than I expected.

> Right.  Dropping this dependency and then deep cleaning the compat code
> is attractive and mostly sidesteps the build system complications.
> That's for a later series.

Yup. Sounds reasonable.

> Ultimately you'd prefer banning mkdtemp(3) instead of automatically
> redirecting to git_mkdtemp(), though, no?

I'm OK either way. Whatever we end up doing for mkstemp(), I think we should match here.

I do like being explicit that we are using our own wrapper and not the system function. But I wonder if it might make complications in third-party code like clar, which calls mkdtemp(). If we don't have a compat macro we'll have to patch the sources that we import into t/unit-tests/clar.

So maybe that is an argument that we should leave the "#define mkdtemp" in place.

-Peff