dimitre
commented
1 month ago I think OF is unreliable today in this state. Libs are still not working OK.
PG is not unreliable for me, today. There were this recent issues difficulties with nlohmann (recently fixed) I know there is the ../ thing you mentioned and I'm glad you are now working on some PR code, so you too can contribute to the closing of this issue here.
Feel free to open one issue for each detail you know it is not working 100% on PG and I can help fixing or discussing what can be done.
In my opinion we can work in the following sequence: Once the libraries are in a stable position (they are not today) we should test all addons and complex projects and different platforms, build systems, make a movement or a "hackaton" as you called before to test And fix PG and templates accordingly if it is missing something. And release 0.13
PS: I'm not using anymore ProjectGenerator frontend (only CLI) in the last 6 years or so, I was using a drag and drop shortcut for the commandline version some while ago and now exclusively ofbuild, so I can make reproducible builds from a yml file
Hi guys, first let me say that I truly and deeply appreciate all of your efforts to make OF's ecosystem a better one, in particular @dimitre and @danoli3 efforts.
Although, PG has become unreliable to me. It worked perfectly well for so long, and while I see that many things need to be updated in order to make it work with newer OSes and IDEs many things broke in the process that did not need to get broken in the first place. Even when I use OF by pulling from github quite frequently, I still need to rely on a quite old version of PG. For me the thing that most bothers me is that local addons no longer work. And at some point I noticed that the parsing addon_config.mk was not done properly and it missed some of the things in it (although I see that got fixed now). Now, I have found that Apothecary is naming what used to be /libs/osx to /libs/macos. I think that that is a really bad move, as I already expressed here. A lot gets broken, for something that looks like an aesthetic change mostly (is there really a technical need for such rename?)
I have already spent several days digging into PG's code and I have to say I am not pleased with what I have found. I see a lot of patched code, some big changes that broke lots (like switching from using strings to use std::filesystem), lots of super redundant code, big chunks of code that are copy/paste, lack of encapsulation, lack of a proper test bed, and the list goes on.
I see that there are some testing scripts but they only test if PG can generate a project for each example but it is not testing if these where properly generated. and still, the examples dont cover the whole use cases (there are many addon's features not being used by the examples) How can we properly test it works? (this would require not only bulding PG but also generating projects that test all the features of PG and have this to compile AND run).
In general I see that since @arturoc left OF's dev lead a lot of its development became messier.
And please don't get me wrong, I am not trying to blame anyone, I am quite sure that all of you do what you do with the best intentions, but we need to be a bit more careful.
What are your thoughts about this @ofTheo @danoli3 @dimitre ?