Oh) Could this be an actual build and just not a dummy step?
Thats adding a lot of building to the serversā¦ Would be better to release a video on how to simply build it yourselfā¦ You can already do it with a bit of try and error through the wiki.
The system might change in the future since Phabricator is not maintained any more.
So things to keep in mind with the new system (they probably already know that)
- better handling with updated revisions of patches to not forget about them
- Accepted patches should not be forgotten to apply (a special row to easily catch that for devs)
- there was also talking about a voting feature for blender devfund supporter but not sure if thats even helpfulā¦
1 Like
Not a big deal I think. Making builds available for non-technical users will dramatically increase number of those, who can try a patch and leave a feedback or a vote. This can become a good guidance for devs on which patch to focus.
Blender devs are overloaded, they canāt dismantle all the patches and write their own new code all at the same time. But at least highlighting highly desired patches can make important patches to land faster.
Why do you need a build with a patch that may not work correctly? Or even damage the files youāre working on. The Blender code changes very quickly and many/most patches cannot be applied without modification. That is, if you canāt build it yourself, then you definitely wonāt be able to adapt patches. Nothing will work automatically.
Anyway, this topic is about checking patches and not about building.
I am still waiting for a case where many people are helpful to make a decision. I get your point and feedback is great but these builds can be outdated pretty fast and building for couple hundred patches a day is not neededā¦ (maybe only specific patches, user can request a build)
Light linking got a lot of attention, but there was never a patch. The baking overhaul. Its not attention or feedback, there are always people who test it. Simply the manpower to review and deliver them to master is not there in many modules. From what i know the hiring process was delayed because of covidā¦
You can even see work on the weekend and so onā¦
1 Like
How about the democracy? =)
You are right, building every single patch every day is not needed. If it can be built once a week - great, but building it just once (at the time of publishing) is probably also good enough. If the patch is outdated and no longer applies - well then, only the last successful build is available.
Cant let that stand thereā¦
You fall into my trap. Democracy is a few deciding for many who voted to represent them.
Anyway what i also wanted to add is that open source projects are different. You get to know every step of development and that can sometimes mean a project is on hold for others like particle nodes. Not a fault of the management there was a long discussion on how to handle it and it was decided to better use that concept in geometry nodes.
Some things can be improved, like a fixed weekly meeting is more beneficial than the current weekly notes in my opinion. Discussion problems and things.
Anyway they can decide, but a better management of patches from the community would be nice.
1 Like
Patch Builds can be done on request and find it here
Your traps made me laugh. Iād suggest to continue our discussion of politics in a private chat. I like that we at least agree, that more feedback from the community is a good thing.
technically it could, but the assumption that all (including future) patches are submitted in good faith, doesnāt seem to be quite in line with the current state of humanity.
5 Likes
I agree itās useless to go on discussing things in circles. Things go wrong (as everywhere) and it would be nice if they didnāt.
However, I donāt want to leave this thread on such a sour note. I have the complete opposite experience. In my experience blender has been one of the more responsive open source projects. I contrast to most open source projects Iāve dealt with all my bugfix patches have been accepted and applied. With new functionality stuff is a lot hairier, but that is a lot more complicated. And lots of community additions have made it into the last few blender releases.
edit: my post was a reply to the exact same post repeated below, which previously was above this one. (adding this edit to keep it understandable)
8 Likes
Blender is being constantly improved by thousands of patches per year contributed by hundreds of people. As long as Blender keeps getting better Iām not at all concerned with how often my own patches get accepted. Best to consider patch submissions as suggestions for changes and not invest too much emotionally in their acceptance; itās not personal.
12 Likes
Youāre naturally free to draw your own conclusions, but I canāt say I agree with them, my position is and has been for a while we can and should do better here, now is everything terrible and the sky falling and donāt bother contributing since nothing will ever happen with your contribution ?
Lets see ! and rather than letting my mood guide this lets back it up with some numbers, I took a look at the last 1000 diffs submitted to d.b.o
| Total | Diffs | Closed | Accpt | Needs Review | Needs Revision | Chg Planned | Abbnt | Draft | |
|---|---|---|---|---|---|---|---|---|---|
| Contributors without commit rights | 90 | 196 | 84 | 1 | 61 | 9 | 2 | 39 | 0 |
| Contributors with commit rights | 55 | 804 | 542 | 28 | 116 | 28 | 15 | 73 | 2 |
| Total | 145 | 1000 | 626 | 29 | 177 | 37 | 17 | 112 | 2 |
or in percentages per group (ie the non committer group had 196 patches, 84 were closed ā 42.86%)
| Total | Diffs | Closed | Accpt | Needs Review | Needs Revision | Chg Planned | Abbnt | Draft | |
|---|---|---|---|---|---|---|---|---|---|
| Contributors without commit rights | 62% | 19.6% | 42.86% | 0.51% | 31.12% | 4.59% | 1.02% | 19.90% | 0.00% |
| Contributors with commit rights | 38% | 80.4% | 67.41% | 3.48% | 14.43% | 3.48% | 1.87% | 9.08% | 0.25% |
Are these numbers great? Iād say we can do better, is the sky falling and weāre neglecting contributors, doesnāt look like it
drilling down further to patches by dev (people without commit rights and 3 or more patches only to keep the data manageable)
| User | Abandoned | Accepted | Changes Planned | Closed | Draft | Needs Review | Needs Revision | Grand Total |
|---|---|---|---|---|---|---|---|---|
| Alaska | 3 | 1 | 5 | 3 | 1 | 13 | ||
| Amudtogal | 1 | 1 | 1 | 3 | ||||
| E-Cycles | 1 | 1 | 1 | 3 | ||||
| erik85* | 5 | 12 | 2 | 3 | 22 | |||
| gfxcoder | 2 | 2 | 4 | |||||
| Gvgeo | 1 | 2 | 3 | |||||
| iyadahmed2001 | 1 | 2 | 3 | |||||
| jarrett.johnson | 2 | 8 | 10 | |||||
| jimman2003 | 2 | 1 | 3 | |||||
| JosefR | 2 | 1 | 3 | |||||
| ktdfly | 2 | 1 | 3 | |||||
| lone_noel | 3 | 3 | 6 | |||||
| Miadim | 1 | 2 | 3 | |||||
| Nikhil-Net | 1 | 2 | 1 | 4 | ||||
| Nutti | 3 | 3 | ||||||
| RedMser | 1 | 2 | 3 | |||||
| t3chstop | 1 | 2 | 3 | |||||
| thibaulltt | 1 | 2 | 1 | 4 | ||||
| tintwotin | 16 | 2 | 18 | |||||
| Wannes | 1 | 4 | 1 | 6 | ||||
| weasel | 9 | 2 | 11 |
- I included Erik85 in this summary, since he gained his commit rights mere days ago, most of his patches in this set of a thousand he had a non commit rights status
I have to admit, you stick out like a sore thumb, but overall, I canāt say Iām super unhappy about these numbers. The people with 1 patch to their name is less encouraging though
| Abandoned | Accepted | Changes Planned | Closed | Draft | Needs Review | Needs Revision | Grand Total |
|---|---|---|---|---|---|---|---|
| 4 | 1 | 0 | 21 | 0 | 24 | 3 | 53 |
24 out of 53 patches still need review, these people have submitted a patch and have seemingly not yet heard back from us, this is why I say āwe can do betterā
Source data used for the tables aboves:
diffs.txt (81.7 KB)
Update: There was some clarifications asked about the statuses the tracker uses
Abandoned - Either reviewer or submitter retracted the patch, could be many reasons, fixed already, not a patch weād want, or as in @tintwotin case the submitter simply walked away
Accepted - Patch has received approval from at least one of the reviewers, it can sometimes stay on this state if multiple reviewers are involved.
Changes Planned - Submitter has changes planned, taking this review out of the review queue for now.
Closed - Patch has landed
Draft - Patch is not quite ready for review yet,
Needs Review - Patch was submitted and is in need of review
Needs Revision - Patch was reviewed and changes were requested.
19 Likes
Best using RCS to do that tho, no?
I think the problem is that most users do not care about side branches enough.
It would be nice to propose different version of blender supervised by others.
for ex @tintwotin why wouldnāt you start supporting your own āBlenderVideo 3.0ā ?
1 Like
No, RCS is a great place for non-programmers to suggest designs for improvement, which could be turned into actual implementations by programmers.
Programmers submit patches as suggested implementations of features, improvements, or bug fixes. We do so voluntarily and they should be treated as gifts to the project to be used - or not - in any way they wish. Itās cool when they use things I give them but Iām not going to get mad when they donāt. Otherwise it is like yelling at a homeless guy because heās not being grateful for the ratty coat you just gave him. LOL
1 Like
wait, am i the homeless guy in this scenario?
1 Like
No, youāre the ratty coat. LOL
1 Like
We do so voluntarily and they should be treated as gifts to the project to be used - or not - in any way they wish.
This statement is not entirely correct. This is true only at the first stage when you add a patch. It can be rejected and there is no problem with that. But when you are asked to change something, then more, then add comments and write a commit message, then this is already a joint work and you expect some final result.
4 Likes
That is a fairly natural expectation, but I personally donāt think like that. For me this is really like volunteering at a soup kitchen or joining a group to clean a local river. Iām using up some spare time trying to make the world a better place, but Iām never going to use time I canāt afford to waste.
3 Likes