2024-11-27T12:31:19+0000 Stuart Henderson <s...@spacehopper.org>: > On 2024/11/27 10:58, Johannes Thyssen Tishman wrote: > > 2024-11-25T13:09:56+0000 Johannes Thyssen Tishman > > <johan...@thyssentishman.com>: > > > 2024-11-24T09:11:44-0800 "Heppler, J. Scott" <shep...@spectrum.net>: > > > > Once I get a -current amd64 system setup, I'd like to start working on > > > > porting FreeCAD. I'm getting library conflicts trying to install KDE 6 > > > > components. I'm anticipating these will resolve in a couple of days. > > > > > > > > Upstream just released a new stable version roughly 11 months after > > > > their prior release. > > > > > > > > https://github.com/freebsd/freebsd-ports/tree/main/cad/freecad > > > > > > > > I think the main dependencies (opencascade, boost, Python311, qt6, GSL, > > > > gtest) are in place. Py_Collada is missing on my ports search but is > > > > optional. It still is a complex port. > > > > > > I have a working port of FreeCAD ready to be submitted. Since I've been > > > working on it for a while, the past couple of days I've been updating > > > the ported dependencies to their last versions and fixing their > > > regression tests. My plan is to submit the port someday this week. If > > > you'd like to test them already, you can find the ports here: > > > https://www.thyssentishman.com/git/ports/log.html > > > > I'd like to submit the port(s) today or tomorrow. In total, I'll be > > submitting 13 ports, so before I send anything I'd like to discuss the > > format of the submission. The port dependency tree is the following: > > > > devel/med > > graphics/pcl > > math/flann > > graphics/py-pivy > > graphics/vtk > > math/cgns > > math/exprtk > > math/verdict > > devel/cli11 > > misc/orocos-kdl > > cad/netgen > > devel/py-test-check > > cad/freecad > > > > In order to make port reviews easier, I'd like to submit all 13 ports > > separately (one email each). To keep track of them, I'd prefix all email > > Please don't, that makes things much harder to review and keep track of > (especially when one port and its required deps are sent separately). > > As long as there isn't anything too complicated, please group a port and > required deps together in a single tar in a single email. e.g. > [graphics/pcl math/flann], or [graphics/vtk math/cgns math/exprtk > math/verdict devel/cli11]. > > (For vtk+deps, i.e. 5 ports, that is about the limit of what you'd want > to send in one go, if it was more than that then splitting the deps in > two would be better). > > If some port is particularly complicated/unusual then please send that > separately first.
Understood. > Try not to have too many separate ports "in flight" at the same time. > > It is however useful to show where the port is in the big picture with > these, nothing complex but just say something like "the ports in this > email are a required dep for X which I have working and will send when > other deps are committed" Got it, I'll try to send them in small groups that make sense to avoid loosing track of them. > > subjects with something like '[FC x/13] ...' and submit them in the > > I can't speak for others, but if I see x/13 all in one batch I will > probably just hit d. Good to know. This pre-submission step has already been worth it then. Thank you Stuart. > > following order to help make sure that dependencies are committed before > > their consumers. > > > > 1. devel/med > > 2. math/flann > > 3. graphics/pcl > > 4. graphics/py-pivy > > 5. math/cgns > > 6. math/exprtk > > 7. math/verdict > > 8. devel/cli11 > > 9. graphics/vtk > > 10. misc/orocos-kdl > > 11. devel/py-test-check > > 12. cad/netgen > > 13. cad/freecad > > > > Does this make sense? Should this be done any differently? > >
