Hi all
Can I assume that the same care is given to the github PR rather then patchset sent to the contributors mailing list?
I have this doubt because I see lack of feedback on GitHub PR (where there is an huge number of PR pending, including two simple from myself) and basically no activity on package PRs. While it seems that the patchset from mailing list submission is a kind of more active (but it may be a personal wrong perception)
Bye
In my experience, it depends on the developers associated with the device and/or code area, their personal time, the complexity of the suggested patch, its perceived value relative to other work, and the perceived risk of implementing the patch.
If you have a patch that you believe to be of significant value to the project in general (in contrast to, for example, fixing a minor bug around a decade-old, underpowered device), you might get more attention with a well-crafted message to the mailing list explaining the problem the patch resolves, the value of the patch, your assessment of the risk, and how that risk has been mitigated.
From my experience simpler git updates are batched together and should get in on a regular schedule. The mailing list seems more a tool for seasoned openwrt devs and feels more closed off/obscure.
I also wait for a big PR for a couple of months, but i understand that's not a simple one. A simple trick that seems to work for openwrt and other projects is to hop into the IRC/discord channel and try a friendly "poke" for a PR, just don't expect a immediate answer, reaction.
Well, I have added support for a new board in sunxi target (Orange Pi PC2) and the THS/DVFS support for all the currently supported Allwinner target (not my job, I have just ported the patch from a well known allwinner source base)
Hauke showed interest in sunxi and he looks over PR's, maybe now is just busy
Ok, but my PR are more than two months old. I mean, maybe it is just a normal time frame, but really I think that there is at least some lack of feedback
@blogic recently stated on IRC that mailing list patches get more attention due to the greater barrier to entry preventing garbage from landing on the mailing list.
Yes, I had to ping him on the IRC to get his attention on a PR that I made 3 weeks ago.
@Menion, I think that sending a patch to the mailing list is the more secure option, because it will be picked by Patchwork, a system which all core developers are familiar with
Then why they don't just shutdown Github PR? Honestly I cannot really understand why all this enjoying of using tool from '90, come on, IRC is even blocked by some ISP!
Github has all the capability to correctly prioritize patch, sanity check them, etc...
1 Like
At least it would be helpful to leave a comment, why some PRs arent merged yet.
I've seen @blogic for example merge some ar71xx device support PRs while other ar71xx PRs (sometimes months old) are ignored. I tried to find a reason he skipped merging them, but I could not come up with an explanation. Maybe he can elaborate, maybe there is no explanation at all.
The TP-Link CPE210v2 PR for example was even submitted twice and the new one is still open after one month. I also stated in the PR that I've tested the PR on my device without any noticeable problem.
I get that all devs are spending their free time on this project and I'm extremely grateful for that. On the other side I can understand that people sending PRs via github get discouraged because they dont get feedback on their PRs.
1 Like
@achterin I agree. On the package github I have just received zero feedback on a very simple, but useful, PRs. On the Openwrt I got feedback on how rework the patch, but then nothing...
Need to poke them on IRC. Or use the mailing list.
GitHub PRs do get merged. They're there for less experienced users that have useful changes.
I reckon the project wants to be open to contributions (and not lose potentially valuable ones). Also, the package feeds seem to rely exclusively on Github nowadays.
There's always the (sometimes inexplicable) tension between people wanting their contribution to be accepted, yet not wanting to go the extra mile to make it easier for developers to review/accept it. That puzzles me. I have no programming background myself, I have basic git knowledge, yet I find it easier to mail my patches rather than open a Github PR. Hanging out on IRC helps, too, btw.
1 Like
I am not questioning on what is the better option, I have my opinion but it is my opinion.
I am just saying that if you give two options, without saying (officially) anywhere that one has a privileged path into the master branch, then both shall be considered the same.
It is pointless that someone does the job in Github (more "complicated" but definitely more structured and modern) and then, after months, discover that shall contact someone in IRC (sick, I don't even know the time zone of the guy!) to ask "please give a look to the PR".
If this is the situation, then just go for one method to push patch so everyone knows what to use
1 Like
I think it's less a matter of one being privileged, as being a matter of
maintainers doing what's easiest for them first.
for some maintainers, responding to github PRs is easier, for others, reviewing
in e-mail is easier.
There is no official preference between the options, both are accepted.