So I looked it up. It is possible to make your own git server and because Abundance runs in the browser If you had a copy of the abundance repo you could run your own abundance server? not planning this just want to know if I understand how it works. I am hesitant to put a lot of design work onto a platform if it can be bricked. Sorry if this is a dumb question, I am just starting to use Github and don’t know how the system works.
Nope all that work is gone. Nothing saved from that day’s work. Will do it again later. I guess it destroyed itself when I was trying to save it? Both the original and the copy I was trying to make are blank. I can step the original back to two days ago so I will do that. Humph.
1 Like
That is really frustrating!
I think that you are right that we will need to support more ways to store things other than just GitHub. GitHub has had horrible reliability issues recently (The Missing GitHub Status Page), but we need to make sure abundance handles it better when github is down so that no data is lost.
Unfortunately, I think this is a issue with replicad and OpenCascade, i’m not totally sure why it’s happning in this particular instance since I’ve mostly seen it when circles intersect with other circles or when their edges are perfectly aligned with a rectangle edge. I’ll keep you updated if i find why this circle is failing in this specific instance.
1 Like
I’m sorry this keeps happening to you. I don’t think it’s autosaving while loading and it shouldn’t be saving while you’re not in the project. I’ll try to find out what’s happening.
1 Like
ok thank you for reporting on this. I will follow up
1 Like
I think it happened while I was in the project in the process of saving where it loaded again. It doesn’t seem to happen when I’m not there. I think it is because my computer loads slowly but the program saves fast. I have fiber internet so that part is good but the computer itself is not fast. One day a new computer. Not now. Thank you for working on the program I am using it knowing there are bugs so that is ok. Just frustrating sometimes.
1 Like
Bar wrote:
I think that you are right that we will need to support more ways to store
things other than just GitHub. GitHub has had horrible reliability issues
recently (The Missing GitHub Status Page), but we need to make sure
abundance handles it better when github is down so that no data is lost.
a big feature of git is that every repo and branch is just as valid as every
other repo and branch.
so instead of all activity going to github immediately, have it go to a local
repo and then sync from there to github as a background task.
this will make things more interesting if you have multiple people modifying the
same design at the same time, but that’s always going to be interesting
the supporting other centralized repos can be added at a later time (gitlab,
there are tools to provide the github api to a git repo you host yourself, etc)
but start by just doing the commit locally and syncing out later.
David Lang
1 Like
Passing through variables Just using variable names seems to be unstable I think it has to do with the order in which things load. In some places they are erased, in others they work.
So I am going back and trying to reconstruct things and strange parts are gone. Mostly direct drawn connections are persistent but some disappeared. Written variables and variable equations inside inputs in rectangles and other shapes disappeared too. I guess mostly what I would like is a way to really save things clearly maybe as a direct download to my own computer in the moment of making things.
Got it rebuilt. Variable names are working ok but sometimes do not pass values up the tree. Numbers in the inputs with those names erase the variable name and reset to 10. When that happens equations inside the molecules maintain Variable names but if one wrote equation script into the inputs of things like rectangles, those get erased. I have replaced all of my shape inputs with equations and direct node wire connections. Oddly Tag nodes also seemed to disappear. Not entirely sure on this one I just noticed broken connections where they were and tag name options were erased. Sorry not better data on that.
It seems like things get lost or erased when different parts of the project calculate and update at different speeds. Big things seem to get lost when the system tries to save while things are loading or when switching contexts, molecules or projects.
I don’t thing these things are because of Github. I was just asking the github questions because I don’t like to have big architecture in projects depend on a company that you can’t control. It seems like it is fine to use them as long as there is a backup plan. Downloading a git repository of the base abundance and then downloading a copy of one’s projects seems fine.
I would suggest: (understanding that this is an opensource vollunteer project)
-
big backup plan, My main concern would be that I don’t understand how the sign in and main screen of abundance works. I don’t want to but would it be possible to run my own Abundance server or is it dependent on the github sign in system?
-
Quick local download option that downloads the project as it is in the browser at that moment with confirmation that it was successful. Maybe a corresponding upload system.
-
Stepback system
-
Stable containers for projects that include all of the assets
-
the ability to copy nodes and node groups from one project to another
-
A way to disable autosave while using a project. This is different than autosave being great when building the project. I don’t know if this makes sense but once I have made a project I would like to be able to open it, and mess with parameters to get a printable result without changing the default project. Maybe the best way is just to make a copy every time you use it but that feels messy especially without file organization
-
a functional delete system for projects. I think I have managed to delete extra copies of things that I make when I am saving obsessively but it has been difficult.
-
Something that prevents Abundance from saving in the middle of loading. Wouldn’t be such an issue if we had a stepback system.
Thank you for your help and for making cool things.
Excited to try to cut this one. This version only has three layers but the actual machine would have multiples of each one.
https://abundance.maslowcnc.com/run/wouldchuckit/Hollander_Beater_Attempt-copy
Tried backing out of it and going back in. It is having trouble finding variables that have been passed into some of molecules but not others. They look ok and still have names in the variable list but maybe are out of sync while loading? Not worried, it still looksl like it is all still there. The message says variable not found. It seems to be variables that I made as an equation in the main layer and then passed values up the tree to subsidiary molecules. I tried retyping the equation or name but no dice yet. Will try again later.
nope, not working. I think i could go throgh it as long as it is an open session and restore all the variables. It works when I am making it but not when it loads. I think variables are not passing numbers into the molecules.
Before Closing
After reopening.
I thought I had made it more stable with wired connections but nope. Will pull all of the lines over and see what happens.
So it appears it is possible to write down an input in the root base layer of a project and then make a molecule and use the name of that variable inside the input of the molecule and it will pass the number one step up the tree of molecules. Apparently it does not work to then make a molecule inside that molecule and pass the value on. Nope. I am doing it with wires from the inputs on the first layer of molecules into the second layer and that fixes it. I was hoping the variable name thing would work but it doesn’t past one level. In this screenshot you can see the ones that I have wired across are now showing values but the ones that are just referenced using words are just sitting at the default 10.
Ok, i have figured out most of it. What is weird is that these equations work in the session where I build them but don’t work after reloading. They seem the most straightforward. I can see the variable input, it has a number in it but the equation isn’t seeing the input? No idea what is going on there. The red text says that it can’t find the variable cover curve, but it is definitely there and spelled the same with a number in it. I’ll change the name of the variable and add a connector wire but not sure what the glitch is in this case.
Also in each section where it is not working it is only one of the variables that I calculated in my main project level and a different one in different places.
If I edit the equation to make it a different variable and then edit it back it works again for that session but then not on reload. I fixed it for me by renaming the variable to a new name and then using a wired connector. That worked even on reloading.
It still recognizes that the variable exists because the equation doesn’t make an input for an unrecognized variable.
This is the current save that I am working inside. https://abundance.maslowcnc.com/wouldchuckit/Hollander_Beater_Attempt-WorkingLayers
@wouldchuck I’m working on going through your posts but in the meantime I wanted you to know i added a disable auto save setting which you might be interested in considering some of the issues you’ve been having with the save. Thanks for reporting. We’re slowly looking through it all.
1 Like
Thank you. I think it will be useful at times. Especially when just opening and poking about with a finished project. I’ll experiment with it when changing projects.
If I have written too much on the other posts I can go back and summarize. I try to document when things go weird but sometimes that gets muddled.
If I get a chance this week I’ll try to make a dummy project that shows the issues
1 Like
Here is a test project that shows the variable issues. It all worked when I wrote it but when I reloaded variables get mixed up. The rectangles in each layer should have variables as their inputs when I made it
https://abundance.maslowcnc.com/run/wouldchuckit/variable_test_project
1 Like
Preview create mode is great. Yay
1 Like
This is extremely helpful. I see your issue with the equation not finding the variable. One thing i see right away and that might simplify your projects is that you don’t need to redeclare the inputs inside every molecule. If you have “a” declared as an input at the top level, you should be able to use that variable inside a nested molecule without it existing at that level. Values for input variables will be evaluated on proximity, if you have an input called “a” in the nested molecule and you use “a” as a parameter value, it will use the closest “a” input it can find, but if it can’t find one in that level, it should keep searching up the tree for other inputs named “a”.
1 Like
I see what’s going on with the input variables. I’ll be working on it. I’ll update you on any progress. Thanks again for making the test project.
1 Like