Rene got further than I did when wishing to contribute something some time ago.
(I just gave up sooner.)
I figured it is what it is, and I don't actually rely on OpenOCD though it's a
handy tool.
John
On Monday, 15 March 2021, 14:11:30 GMT, Tommy Murphy
<tommy_mur...@hotmail.com> wrote:
Thanks a lot Rene.Very interesting to get your take on it.Thanks for taking
the time to summarize it so clearly.
From: Rene Kita <open...@rkta.de>
Sent: Monday 15 March 2021, 10:22
To: openocd-devel@lists.sourceforge.net
Subject: Re: [OpenOCD-devel] [RFC PATCH] Remove warnings
On Sun, Mar 14, 2021 at 09:38:06PM +0000, Tommy Murphy wrote:
> > You make it really complicated to submit minor patches.
>
> I'm curious to know what you think is really complicated about the process?
> I haven't been through it myself yet but it looks reasonable to me.
> What specifically causes you to give up?
> Thanks
To clarify this: This was a small patch and I don't expect to contribute
any further patches in the foreseeable future. That's why I send it to the
ML as [RFC] and explicitly said the maintainers may consider the patch a
bug report. My intent was to make this as less time consuming as
possible for everybody involved[1].
BTW, Antonio Borneo took care of it and added it to Gerrit with a fixed
commit messages and attribution - thanks!
tl;dr: I took me way more time to find out what I need to do to submit
the patch to Gerrit than I needed to write the patch send it to the
mailing list. Also the process leaves me with an setup (account,
ssh-key) which I don't need and don't want to maintain.
How it went after I was asked to submit it to Gerrit:
I open up Gerrit only to get a message that I need JavaScript, so I open
it in another browser. I click on register, go to the first text field
and enter my email address, press enter - error message. Now I see that
the button below the field is named 'Sign In'. I go back to make sure I
actually clicked the 'Register' link and not accidentally 'Sign In'.
Same page - OK, maybe somewhere below...
I see some OAuth-Options which I do not intent to use (It's non of
Googles or GitHubs or whomever business what other services I use -
especially not to get noticed when ever I use them). So I click on
'Get OpenID' as it is the last option left. I'm greeted with a lot of
text how to contribute to the OpenID community - I only want to submit
this patch and I don't even know what OpenID is, WTF!
At this point I stopped caring and closed the browser, deciding to tell
the devs to take the patch or leave it - I don't care.
After a good night of sleep I'm calmed down and decided to give it
another chance. I have an GitHub account, so I could use it if
absolutely necessary. I use OpenOCD daily, I need it for my work, it's a
great tool - if the maintainers insist to not take patches via mail I'll
show some effort to submit the patch via their preferred way.
I click on the GH OAuth link to be greeted with a lot of warnings and
'Are you sure...'s. I read about the security implications to make sure
that I indeed are sure what I'm doing. My GitHub Account is linked to my
company's organization and I don't want to risk to do something stupid.
I'm finally decided that's OK to give permission and are logged in to
Gerrit. Looking for a way to upload a file / add file which I don't
find. I said OK, obviously I really need to read these 242 lines of
HACKING.
There I found I need to add an ssh key and install some git hooks.
And that was the point where finally send this email stating that it is
to complicated to send a minor patch.
And now I have a Gerrit account that I can't delete and don't know what
happens if I revert the token from Github. Which is another reason why I
hesitated to create one.
I understand that a project has to have defined workflows and I have to
adept to them when I want to work with this project. But this was a
one-shot patch. I don't even need attribution for it. I fixed it,
because I use a experimental build of GCC and wanted to save the project
from build failures (-Werror is on by default) in the future. I also
don't see major contribution from myself because everything I use
OpenOCD for works already. And as I only use it for STM32 and ST is a
contributor using OpenOCD for their own IDE I expect it stay this way.
I'm sure there was some discussion that let to the decision not to
accept patches via the ML that I'm not aware of and I'm certainly not in
the position to question it or restart the discussion - but to bring a
solution and not problems: I suggest to be less strict with minor
patches.
[1] And now I finish this footnote nearly an hour after I started
writing this mail. :)
_______________________________________________
OpenOCD-devel mailing list
OpenOCD-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openocd-devel
_______________________________________________
OpenOCD-devel mailing list
OpenOCD-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openocd-devel
_______________________________________________
OpenOCD-devel mailing list
OpenOCD-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openocd-devel
