On 11/13/2016 1:47 PM, Nick Østergaard wrote: > 2016-11-13 19:29 GMT+01:00 Wayne Stambaugh <stambau...@gmail.com>: >> On 11/13/2016 1:15 PM, Nick Østergaard wrote: >>> 2016-11-13 18:59 GMT+01:00 Wayne Stambaugh <stambau...@gmail.com>: >>>> Hey Nick, >>>> >>>> The problem with the boost 1.61 bug fix in the master branch is that it >>>> includes c++11 code. That's why I didn't merge it into the stable >>>> branch. I want to avoid c++11 in the stable 4 branch so we will have to >>>> let package devs know that they cannot build kicad against boost >= >>>> 1.61. If someone can fix the boost context issue without using any >>>> c++11 code, I would accept a patch for that. >>> >>> Good. >>> >>>> >>>> I specifically changed the version string code to accept only two >>>> conditions for creating version strings. Case 1 is using git to >>>> generate the version string when KicadVersion.cmake is set to the >>>> default which it is current. Case 2 is that I (or the release manager) >>>> sets the version and branch settings in KiCadVersion.cmake which >>>> overrides the git version string. I don't want developers setting this >>>> as we have in the past. >>> >>> What do you mean by developers setting this? I don't think any >>> developers have set anything here. >> >> Prior to the changes I made, you had to set KICAD_BUILD_VERSION and/or >> KICAD_REPO to generate a stable version string. Otherwise, you either >> go the git version string or no-vcs-found. This didn't work for package >> devs who expect the source archive to build with the appropiate version >> string without defining them at config time. >> >>> >>>> Apparently package devs hate it because they >>>> had remember to set the version and branch strings at config time when >>>> building from source archives otherwise they end up with no-vcs-found >>>> version string. >>> >>> Well, I have a hard time to see how the current solution in 4.0 is >>> addressing this issue. Does this mean that on the commit where you tag >>> it 4.0.5 will change the set( _wvh_version_str "no-vcs-found" ) with >>> set( _wvh_version_str "4.0.5" )? >> >> Once the version is set to something other than "no-vcs-found" in >> KiCadVersion.cmake, it takes precedence. What I plan on doing is >> setting it to 4.0.5 in the stable version tagged as 4.0.5 (hey that >> almost makes sense!). When I create the source archive, it wont be >> necessary to define KICAD_BUILD_VERSION or KICAD_REPO. Once that is >> complete, I will can change the version string in KiCadVersion.cmake to >> "4.0.6-pre" or something like that. >> >>>> I see Chris added this back into the master branch but >>>> I'm not sure I'm going to keep it. Developers should be using the >>>> version string generated by git and package devs should be using the >>>> version set in KiCadVersion.cmake. I can't see any reason for a third >>>> option. >>> >>> As it works in 4.0 now, I can't set -DKICAD_BUILD_VERSION=4.0.5_rc1. >>> It will still use the git version. >>> >>> I would expect to be able to override the version string in any case. >>> For example if I wanted to make a release candidate build to sort out >>> some packaging issues, it would be nice to see from the build that it >>> was a release candidate if anyone reports a bug on it. >> >> If you really need to be able to define the version at config time, I >> can leave Chris' recent changes assuming they didn't break the behavior >> of KiCadVersion.cmake. Do other projects allow changing the build >> version at config time? I got the feeling from some of the comments >> that this was "atypical". > > I am ok with it working the other way where we can have a hardcoded > version in the source for the stable branch. But really, I think it > should be set now as 4.0.5-devel then. Although I do not have any hard > feeling for using, -pre, -dev, -devel or whatever that is not -rcX. > > But I think it should always be overridable by defining it manually. > For eample for the nigthly PPA's they have a version string like > "201611122049+7349~55~ubuntu14.04.1" and to track down issues with > that if you want to know what exact build and source is used it is > easier to find if that could be used a version string. > > If this happens, I am also ok with removing the KICAD_REPO_NAME option > and only use KICAD_BUILD_VERSION to override. > > I hope you can understand why I still want it to be overridable.
I'm fine with that. I guess I should have asked before I removed it. The order of precedence should be: configuration defined, KiCadVersion.cmake, then the git version. @Chris, is this what your changes do? I didn't get time to look at them and this week is going to be crazy at work so I may not have time to look over them until next weekend. > >> >>> >>>> >>>> I will be in an out all day today so if I don't reply I'm not ignoring >>>> you I've just got a full plate today and will respond as soon as I get a >>>> chance. >>>> >>>> Cheers, >>>> >>>> Wayne >>>> >>>> On 11/12/2016 11:55 AM, Nick Østergaard wrote: >>>>> Hi Wayne, >>>>> >>>>> Good news to see a new release, but I wonder about two things: >>>>> >>>>> 1. Boost 1.61 compatibility and bugfix? >>>>> 2. Release version string. >>>>> >>>>> Should the release should include the boost 1.61 fix. It was discussed in: >>>>> [Kicad-developers] About Bug 1604841: Pcbnew crash at moving via, and >>>>> boost::context fixes to make it compatible with boost 1.61 >>>>> >>>>> And there is a patch in [1], although it does not explicitly state >>>>> where it is from, but it looks similar to [2] and [3]. With out the >>>>> patch in [1] I can not build it against boost 1.62 on archlinux. >>>>> >>>>> Secondly, I would like to know how we are supposed to set the release >>>>> infor for the version string. For the 4.0.0--4.0.4 you are supposed to >>>>> use the following options to set the version string: >>>>> >>>>> -DKICAD_REPO_NAME=stable -DKICAD_BUILD_VERSION=4.0.5 >>>>> >>>>> But KICAD_REPO_NAME does not seem to exist anymore... even though I >>>>> set KICAD_BUILD_VERSION it gets overridden by the git info and the >>>>> version string becomes: >>>>> #define KICAD_FULL_VERSION "(2016-11-09 revision 18f77b8)-4.0" >>>>> >>>>> I would have expected to get something like we are used to as, >>>>> "4.0.5-stable" for example. >>>>> >>>>> Nick >>>>> >>>>> [1] >>>>> https://git.archlinux.org/svntogit/community.git/tree/trunk/boost-1.61.patch?h=packages/kicad >>>>> [2] >>>>> https://git.launchpad.net/kicad/commit/?id=06d4894fdbeb00727cdcc667b8899ad73d8eb1c2 >>>>> [3] >>>>> https://git.launchpad.net/kicad/commit/?id=78bc3c65de6c03d19be9902327d08cd4d87c229c >>>>> [4] https://lists.launchpad.net/kicad-developers/msg25556.html >>>>> >>>>> 2016-11-11 23:44 GMT+01:00 Wayne Stambaugh <stambau...@gmail.com>: >>>>>> This is the last call for patches to the 4 stable branch. If I don't >>>>>> hear anything, I will set version string and tag 4.0.5 in the stable >>>>>> branch on Sunday and give our library, doc, translation, and package >>>>>> devs a couple of weeks to prepare before I make the official >>>>>> announcement. >>>>>> >>>>>> Thanks, >>>>>> >>>>>> Wayne >>>>>> >>>>>> _______________________________________________ >>>>>> Mailing list: https://launchpad.net/~kicad-developers >>>>>> Post to : kicad-developers@lists.launchpad.net >>>>>> Unsubscribe : https://launchpad.net/~kicad-developers >>>>>> More help : https://help.launchpad.net/ListHelp >>>> >> _______________________________________________ Mailing list: https://launchpad.net/~kicad-developers Post to : kicad-developers@lists.launchpad.net Unsubscribe : https://launchpad.net/~kicad-developers More help : https://help.launchpad.net/ListHelp