Bert, you beat me to the punch! I have a draft email = open right now that mentions some of these things.

I thin= k that most of the commits in my 4.0.0 branch should probably be in master = as well.

The MinMaskGap failure is the result = of a non-zero diff between the golden file and the pcb output file. The non= -zero part is because the version strings that pcb writes at the beginning = of the file are different. The test isn't really failing. If you update= the version string in the golden file to match the new version, then the t= est will pass.

I tried to create a 4.0.0 branch in the gi= t root, but was not able to. Maybe I don't have the requisite permissio= ns?

I've been able to successfully do a make distchec= k with each of the hids (gtk, lesstif, and batch) on my Mac. I was just wor= king on doing the same on a Slackware VM and a Debian VM, but am running in= to some other problems there.

--Chad

On Fri,= Dec 30, 2016 at 5:27 PM, Bert Timmerman (bert DOT timmerman AT xs4all DOT nl) [via geda-user AT delorie DOT com] <geda-user AT delorie DOT com><= /span> wrote:

Chad Parke= r (parker.cha= rles AT gmail DOT com) [via geda-user AT delorie DOT com] wrote:

Is the reason people want to skip to version 4 just because of that fork th= at claimed version 3.x? If so, we could still use the 2.x.x series before h= aving to worry about the naming collision, and then skip to 4.0.0. Or is it= that we want to tell prospective users that ours is better by having a hig= her version number?

DJ-
I agree that there haven't been any changes that would break backward c= ompatibility, but sometimes you just have to make a clean break, like when = the kernel went to 3.0.0.

Bert-
Thanks, I can be a little OCD sometimes, and I do like things to be neat an= d clean! Particularly when it comes to plans.

I sort of assumed that we would either have the code sprint earlier, or not= have one this month since the last Sunday of the month (tomorrow) is Chris= tmas day and I imagine that a lot of people will be observing that holiday.=

I'll run through the release procedure that Dan pointed out sometime in= the next couple of days and let folks know how things go. I think there ar= e at least a couple of compile warnings on my Mac that will need to be fixe= d.

Cheers, and have a good holiday to all those who are celebrating.
--Chad

On Sat, Dec 24, 2016 at 3:01 PM, Girvin Herr (girvin_herr+gherr_lists AT fast= mail.com <mailto:girvin_herr%2Bgherr_lists AT fastmail DOT com= >) [via g= eda-user AT delorie DOT com <mailto:geda-user AT delorie DOT com>] <geda-user AT delorie DOT com <m= ailto:geda-user@= delorie.com>> wrote:

=C2=A0 =C2=A0 On 12/24/2016 04:50 AM, M. J. Everitt (m DOT j DOT everitt AT iee DOT org
=C2=A0 =C2=A0 <mailto:m DOT j DOT everitt AT iee DOT org>) [via geda-user AT delorie DOT com
=C2=A0 =C2=A0 <mailto:geda-user AT delorie DOT com>] wrote:

=C2=A0 =C2=A0 As a prospective proxy maintainer for Gentoo, please let me k= now
=C2=A0 =C2=A0 when you roll a tarball, and I'll run through the Gentoo<= br> =C2=A0 =C2=A0 packaging process, and do some preliminary build/run-testing = and
=C2=A0 =C2=A0 feed back.

=C2=A0 =C2=A0 If wanted, I can do the same for a Slackware Linux package.= =C2=A0 I am
=C2=A0 =C2=A0 currently running Slackware 14.1 (k3.10.103), but I have a co= py of
=C2=A0 =C2=A0 the latest 14.2 and plan to install it early next year.=C2=A0= The only
=C2=A0 =C2=A0 change I can see is that the package architecture went from i= 486
=C2=A0 =C2=A0 in 14.1 to i586 in 14.2, but that is trivial.=C2=A0 Of course= , the
=C2=A0 =C2=A0 dependencies have changed.

=C2=A0 =C2=A0 Another option would be to contact the SlackBuilds.org pcb pa= ckage
=C2=A0 =C2=A0 maintainer, Felix Pfeifer:=C2=A0 pfeifer [dot] felix [at] goo= glemail
=C2=A0 =C2=A0 [dot] com
=C2=A0 =C2=A0 He might be willing to do this task and put the resulting pac= kages
=C2=A0 =C2=A0 on SlackBuilds.

=C2=A0 =C2=A0 geda-gaf on SlackBuilds.org seems to be maintained by a diffe= rent
=C2=A0 =C2=A0 maintainer, Stephen Van Berg:=C2=A0 stephen_van_berg at earli= cker dot com

=C2=A0 =C2=A0 I would second a more consistent version numbering system, as=
=C2=A0 =C2=A0 this makes version-tracking easier for packagers too, so pcb = v4
=C2=A0 =C2=A0 would indeed be a good choice if settled upon!

=C2=A0 =C2=A0 Consistency, yes.=C2=A0 But the form doesn't matter for S= lackware
=C2=A0 =C2=A0 packaging.=C2=A0 x.y.z is just as fine as yyyymmdd.=C2=A0 The= only caveat is
=C2=A0 =C2=A0 that dashes may not be used in the version string.=C2=A0 I us= ually
=C2=A0 =C2=A0 change them to underscores or remove them entirely if in yyyy= -mm-dd.

=C2=A0 =C2=A0 Girvin Herr

=C2=A0 =C2=A0 Good luck, thanks, and keep us posted on the list :)

=C2=A0 =C2=A0 Michael.

=C2=A0 =C2=A0 On 24/12/16 12:23, Peter Clifton (petercjclifton AT googlemail DOT com
=C2=A0 =C2=A0 <mailto:petercjclifton AT googlemail.com>) [via
=C2=A0 =C2=A0 ge= da-user AT delorie DOT com <mailto:geda-user AT delorie DOT com>] wrote:

=C2=A0 =C2=A0 Nice to hear work is going on with mainline pcb as well...
=C2=A0 =C2=A0 One thing prior to release. We ought to git rm the debian
=C2=A0 =C2=A0 directory that added to the repository.

=C2=A0 =C2=A0 This matches preferred distro policy, and makes life easier f= or
=C2=A0 =C2=A0 the official packagers (so they can use our sources directly<= br> =C2=A0 =C2=A0 without having to repackage them to remove the debian dir.
=C2=A0 =C2=A0 I checked this with our debian packagers some while back to =C2=A0 =C2=A0 confirm, but never got around to making the delete.

=C2=A0 =C2=A0 Peter

Hi Chad,

I see you are busy with the 4.0.0. test drive in your "home/cparker/4.= 0.0" branch.

AFAICT some of those commits should be in master, others should go in a &qu= ot;pcb-4.0.0" release branch once you have finished testing.

One other thing is that I found is that "make check" fails on the= MinMaskGap test ... in a pristine Ubuntu 12.04 (precise) environment, see = https://travis-ci.org/bert/pcb/builds and https://bugs.launchpad.net/pcb/+bug/1653280.

I think we need to squash this bug before we can ship 4.0.0 ("make che= ck" failing can be seen as "bad form").

I will keep you informed about any findings in the bug report.

Kind regards,

Bert Timmerman.

--001a114435b8bdfe990544e8dc08--