-
Notifications
You must be signed in to change notification settings - Fork 45
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Lack of Feedback from Maintainer #80
Comments
I think @mswart is fine with us merging new parsers and fixing existing ones (I've done so multiple times over that last year). I usually try to respond quickly to new pull requests, but I am currently working on my master thesis so I haven't had much time lately. I intend to invest more time once that is done (June/July). Then I will also come back to your pull request regarding regression testing. |
There is also the other repos (for PyOpenMensa and OpenMensa) that have PRs that are untouched for months now. I just felt like here are more people to discuss this with. :) Maybe a "democratic" approach like we have here might work. But I think then it might be helpful to have some guidelines on what to look for when merging new parser etc, since I personally am a bit unsure of how to judge PRs and existing parsers. What do you think of aggregating rules/things to look for in a guideline document? |
There is also the issue of diffusion of responsibility. Since I'm unsure of how to deal with issues/PRs, and I'm not the only one with merge rights, I just lay back and let you guys deal with it. :) Having some "roles" helps with this, since I know who to ping/ask/replace for a given task. |
I just reviewed the saarland parser pull request, if you would like some inspiration 😉. I generally go through the code line by line and have the source (api or website) and the generated xml open for reference. This way I "manually" parse the source in my head. If something is unclear or wrong, I add a comment. I also usually make some general notes separately, that get condensed into one or more comments at the end. Sometimes I play around with the code a little bit to test robustness and edge cases.
I am not sure if we need roles, but feel free to ping me directly if you are unsure about anything. |
Thanks, I will try to have a look at it later on. ;) Is there a way to send you a PM? If not, maybe it's a good idea to start a Discord or similar channel for quick and dirty communication and questions? |
Just send me an email (address is on my profile). 😉 |
Ok. :D So easy. What do think of a Discord/Gitter, though? |
FYI: #89 |
@klemens, I don't know where else to write this: Thank you! |
I am going to close this. The most realistic path forward is probably moving the parsers and the python lib into the openmensa org, but like mentioned in another thread, I currently don't have time for that. |
I just wanted to discuss the issue of the frankly slow feedback and merges from @mswart. It's difficult to stay excited and help with this project, without getting meaningful feedback and changes.
I know it sounds harsh, but I do not mean it as a personal attack. He has built all of this and continues to run everything, and I am sure everyone is very thankful for all the effort he puts into this project.
This issue is primarily about discussing how we all can help with the fact that @mswart has other responsibilities than this project. None of us, can spend all their time on here, which makes the contributions and work from @mswart even more impressive.
Maybe there is some simple change that will help solve this issue. Could we declare a "vice-president' that has the authority to guide the project when he's inactive? Maybe @mswart can check in twice a week, without sacrificing too much time?
I'm just throwing ideas around. What I'd like is to discuss this issue, because I truly believe that it hurts OpenMensa.
Again, @mswart is at no fault. I know that it easily sounds like I would think that. Changes in the amount of time someone can spend on a project like this just doesn't stay the same forever. This is just about figuring out how we can best deal with this. :)
The text was updated successfully, but these errors were encountered: