<teepee>
Hmm, the CGAL exception is probably not needed anymore, not sure it can be removed
la1yv_a has quit [Remote host closed the connection]
la1yv_a has joined #openscad
ferdna has joined #openscad
ur5us has joined #openscad
ferdna has quit [Quit: Leaving]
<joseph_>
InPhase: Could you please clarify what your last message meant by "extending the shader interface"? Do you mean how to support future changes to the OpenSCAD UI, or the definitions of classes in the codebase?
<InPhase>
joseph_: Which I interpret as, when we are going to do this in general for submitted and not hard coded with one entry, the interface needs to be designed well in a way that is well-specified, reasonably consistent over time, but also considerate of future needs.
<InPhase>
joseph_: And there seem to be options for doing this more elegantly than we have.
<peepsalot>
joseph_: as InPhase mentioned, /usr/share/openscad is not going to be writable for normal users. If a GUI driven method of adding shaders is wanted, then the destination should probably be alongside the user library path, eg: ~/.local/share/OpenSCAD on Linux
Chrischl has joined #openscad
<joseph_>
InPhase: What if I describe in the proposal that at the backend, a new class would serve as a single source for loading shaders based on user selection, and then passing the results to Render.cc? Would you suggest any changes to how I've described the UI organization ?
linext has quit [Quit: Leaving]
<peepsalot>
I think the extension to shader interface refers to API, such as supported GL Uniforms and Vertex Attributes. some shaders may require more or different forms of data about the scene
<peepsalot>
I have some code I wrote a while back which defined a set of common uniforms and attributes that I could see shaders possibly using
<peepsalot>
unfortunately it never got merged and is from before the great reformatting/restructuring and I have been struggling with git's inability to track renamed files to get things up to date with master
<peepsalot>
joseph_: There are a few GL commands which can be used to check the names and number of GL Uniforms and Attributes used by a shader. As long as they stick to a common set of possible names, then OpenSCAD can support a wide variety of cases
<peepsalot>
glGetActiveAttrib and glGetActiveUniform
othx has quit [Ping timeout: 256 seconds]
othx has joined #openscad
juri_ has quit [Ping timeout: 250 seconds]
juri__ has joined #openscad
juri__ has quit [Read error: Connection reset by peer]
Xeha has joined #openscad
Joel has quit [Ping timeout: 248 seconds]
juri_ has joined #openscad
Joel has joined #openscad
<joseph_>
peepsalot: I will add a note about those to my proposal, thank you. In order to recover the old files, is it possible that the history could be untangled by tuning Git's similarity index used for finding renames? Or am I not understanding the issue fully?
<peepsalot>
looking over my old code, i would hate to have it completely go to waste. i hope it can be used as a base for GSoC shader changes. i've been overly fatigued lately though, and find it hard to get much accomplished or to stay motivated
<peepsalot>
joseph_: the similarity % can be tuned, but i could not find a value that fixed all false deletions, without going too low and introducing false matches between unrelated files.
Chrischl has quit [Ping timeout: 250 seconds]
<peepsalot>
the problem is that there was a large reformating that affected the majority of lines of code, making consistent indentation across the project. then a separate commit shortly after which organized the src tree into subdirectories. those two in combination result in a git merge nightmare for any branches made before them
<InPhase>
peepsalot: If the changes were just in a few files, the simplest is probably just to extract and gist them as a reference.
<InPhase>
Perhaps I am quick to suggest working around rather than with git at times... but there are reasons I leap to this so often. :)
RichardPotthoff has joined #openscad
<peepsalot>
InPhase: 32 files have modifications
<InPhase>
Well that's unfortunate.
<InPhase>
peepsalot: The project work period start date is June 13th, which is approximately 2 months out. I would interpret that as there's no rush to try to hammer out that code extraction while you're still feeling fatigued. You have time to get a sharper more energetic moment.
<InPhase>
peepsalot: Maybe this applies to you as well, but I realized over the past 2 years I had been radically underutilizing my PTO days, to the point that they had saturated out and stopped accumulating. Some of my critical milestones have passed though, and I'm making an effort to balance my life out more. Probably, realistically, I should have tried to do this before those core milestones were completed,
<InPhase>
because I was often not at peak operating capacity anyway after pushing too hard.
<peepsalot>
i've been unemployed for a while now. :( I actually need to get a job very soon as my funds are running low. i think it may be sleep apnea but have not been diagnosed
<InPhase>
Ah, yeah, different sort of problem then.
<InPhase>
Are you trying to stay in your area for employment?
<peepsalot>
yes, ideally
ur5us has quit [Ping timeout: 250 seconds]
<joseph_>
InPhase: Agreed, there's no rush to recover these changes now. peepsalot good luck with the job search and sleep issues, those are much higher priorities. All I need to know now is any remaining feedback about my proposal. I will submit tonight and edit later if needed
retrosenator has quit [Ping timeout: 246 seconds]
ferdna has joined #openscad
retrosenator has joined #openscad
<InPhase>
joseph_: Try to get something with good depth in about the interfaces, such as with some plans about the parts you can figure out now with a little investigating, and a specified plan for how you will figure out other things that you don't yet know about it. That will illustrate some realistic thinking about some of the practical complexities that will arise.