Date: Wed, 16 Aug 2023 12:11:32 -0600 From: Warner Losh <imp@bsdimp.com> To: FreeBSD Current <freebsd-current@freebsd.org> Subject: HEADS UP: $FreeBSD$ Removed from main Message-ID: <CANCZdfoe-HHimYsO5socTXqr20qvZSHoLou7ede2SpcvZ2aLow@mail.gmail.com>
next in thread | raw e-mail | index | archive | help
--000000000000d1017b06030e390a Content-Type: text/plain; charset="UTF-8" Greetings, I've just pushed the results that remove 31,035 instances of $FreeBSD$ from the main branch. >From this point out, we are effectively done with $FreeBSD$, though there's 4 places where $FreeBSD$ still exists in the tree: (1) In the contrib area, we have 621 remaining. I didn't remove any from contrib code. (2) There's 140 $FreeBSD$ or similar tags part of 'static const char rcsid[] = "$FreeBSD$". There's too many different styles to get with automation, and they are also intertwingled with sccsid[] entries, other SCCS tags and copyright strings. These will be sorted out separately, since we need to talk about old sccs tags in the tree and sort that out too at the same time. (3) There's one $FreeBSD$ in a version tag for bootinfo for chrp bootinfo.txt. I don't know the effects of removing this entry, so I left it in place. Not sure it's worth fixing. (4) indent tests remove $FreeBSD$ so that it doesn't screw them up. This is likely harmless, but could be removed. I didn't want to mess with it, though. I've removed 99.5%+ of the 'live' instances in the tree. The ones from contrib should be removed upstream and I plan to do a MFC-like thing where I cherry-pick some commits, and run my script against stable/13 and include the main hash so our scripting things it's a real MFC (the diff hashes will differ, so git's native tooling will take the slow path, at least, and may get confused otherwise. I've built world on aarch64 and amd64. I've built kernels on all the architectures after this change... or well, last night's main. I eye-balled today's changes and they all look good, but there's no incremental building with this change, so I'm starting a new universe after I loop the change back into my tree. Also: expect long build times, git fetch times, etc after this. Thanks to the many people who gave me feedback on the details of this change and how to chunk it up. Hopefully the 40ish commits was the right balance between 'all at once' and 'every dir'. As always, if there's any problems after this change, please let me know. Warner --000000000000d1017b06030e390a Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr"><div>Greetings,</div><div><br></div><div>I've just pus= hed the results that remove 31,035 instances of $FreeBSD$ from the main bra= nch.</div><div><br></div><div>From this point out, we are effectively done = with $FreeBSD$, though there's 4 places where $FreeBSD$ still exists in= the tree:</div><div>(1) In the contrib area, we have 621 remaining. I didn= 't remove any from contrib code.</div><div>(2) There's 140 $FreeBSD= $ or similar tags part of 'static const char rcsid[] =3D "$FreeBSD= $". There's too many different styles to get with automation, and = they are also intertwingled with sccsid[] entries, other SCCS tags and copy= right strings. These will be sorted out separately, since</div><div>we need= to talk about old sccs tags in the tree and sort that out too at the same = time.<br></div><div>(3) There's one $FreeBSD$ in a version tag for boot= info for chrp bootinfo.txt. I don't know the effects of removing this e= ntry, so I left it in place. Not sure it's worth fixing.<br></div><div>= (4) indent tests remove $FreeBSD$ so that it doesn't screw them up. Thi= s is likely harmless, but could be removed. I didn't want to mess with = it, though.</div><div><br></div><div>I've removed 99.5%+ of the 'li= ve' instances in the tree. The ones from contrib should be removed upst= ream and <br></div><div><br></div><div>I plan to do a MFC-like thing where = I cherry-pick some commits, and run my script against stable/13 and include= the main hash so our scripting things it's a real MFC (the diff hashes= will differ, so git's native tooling will take the slow path, at least= , and may get confused otherwise.</div><div><br></div><div>I've built w= orld on aarch64 and amd64. I've built kernels on all the architectures = after this change... or well, last night's main. I eye-balled today'= ;s changes and they all look good, but there's no incremental building = with this change, so I'm starting a new universe after I loop the chang= e back into my tree. Also: expect long build times, git fetch times, etc af= ter this.</div><div><br></div><div>Thanks to the many people who gave me fe= edback on the details of this change and how to chunk it up. Hopefully the = 40ish commits was the right balance between 'all at once' and '= every dir'.</div><div><br></div><div>As always, if there's any prob= lems after this change, please let me know.</div><div><br></div><div>Warner= <br></div></div> --000000000000d1017b06030e390a--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfoe-HHimYsO5socTXqr20qvZSHoLou7ede2SpcvZ2aLow>