Hi Jan, > The main motivation was considerable reduction of build time of our > whole codebase.
If you want to optimize build time, I recommend giving [1] a try. On all five platforms and devices I've tested - including Windows VMs - it cuts build time in half, at the very least. I've rebased the patch to the current master to make it easier for you to test. I'd appreciate any feedback. [1] https://review.openocd.org/c/openocd/+/9541 Best regards, Marc On Tue, 2026-05-26 at 07:03 +0000, Jan Matyáš wrote: > > Hi Paul, > > > > Also interesting what makes you use clang on windows, did any of > > major > > environments (such as Cygwin or MSYS2) switch or is there a new > > great > > one? > > > Within Codasip, we have recently switched from GCC to Clang for > Windows builds. We are using Clang distributed in MSYS2. > The main motivation was considerable reduction of build time of our > whole codebase. > > > Regards, > Jan > > > > > > > > From: Paul Fertser <[email protected]> > Sent: Monday, May 25, 2026 3:09 PM > To: Jan Matyáš <[email protected]> > Cc: Marc Schink <[email protected]>; OpenOCD > <[email protected]> > Subject: Re: OpenOCD: last call for v1.0.0-rc1 > > > > > Hey Jan, > > On Mon, May 25, 2026 at 06:36:43AM +0000, Jan Matyáš wrote: > > I also recommend to add Clang Windows build job to Jenkins. Who has > > the > > permissions to do that? How can I help to make this happen? > > No special help needed but that will increase the time each and every > CI build takes, so a consensus needs to be reached first. > > Also interesting what makes you use clang on windows, did any of > major > environments (such as Cygwin or MSYS2) switch or is there a new great > one? > > I wonder if it's a good opportunity to revise the test matrix?
