2022-02-26 01:51:32

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v2 0/7] clk: Declare mux tables as const u32[]

Quoting Jonathan Neuschäfer (2022-02-05 02:36:06)
> I noticed that the 'table' parameter to clk_register_mux_table is never
> used for modifying the table elements, and so it can be declared const.
>
> In version 2 I'm addressing two warnings in the clk-lpc18xx-cgu driver
> that I previously missed.

The format of these patches deeply confused my scripts that use git
interpret-trailer. I fixed it now, hopefully it keeps working. In the
future, please don't add more triple dash '---' sections to the patch.
It looks like those extra sections for the changelog messed everything
up. Or there's a new bug in interpret-trailers. Either way,
interpret-trailers was adding tags after the entire patch contents
because I think it looks for the last triple dash instead of the first
triple dash. Not sure why it's done that way. I resorted to
reconstructing the patch after splitting it with mailinfo.


2022-02-26 20:08:18

by J. Neuschäfer

[permalink] [raw]
Subject: Re: [PATCH v2 0/7] clk: Declare mux tables as const u32[]

On Fri, Feb 25, 2022 at 05:00:31PM -0800, Stephen Boyd wrote:
> Quoting Jonathan Neuschäfer (2022-02-05 02:36:06)
> > I noticed that the 'table' parameter to clk_register_mux_table is never
> > used for modifying the table elements, and so it can be declared const.
> >
> > In version 2 I'm addressing two warnings in the clk-lpc18xx-cgu driver
> > that I previously missed.
>
> The format of these patches deeply confused my scripts that use git
> interpret-trailer. I fixed it now, hopefully it keeps working. In the
> future, please don't add more triple dash '---' sections to the patch.
> It looks like those extra sections for the changelog messed everything
> up. Or there's a new bug in interpret-trailers. Either way,
> interpret-trailers was adding tags after the entire patch contents
> because I think it looks for the last triple dash instead of the first
> triple dash. Not sure why it's done that way. I resorted to
> reconstructing the patch after splitting it with mailinfo.

Hmm, sorry about that.

I've used this format for a while, because it conveniently lets me
keep my remarks in a git commit (rather than a patch file), until I use
git format-patch to generate the final patch file.

I'm not very familiar with git interpret-trailers, but git 2.34.1
doesn't seem to get confused on my patches (or I didn't pass the right
options to cause it to happen).


Jonathan


Attachments:
(No filename) (1.39 kB)
signature.asc (849.00 B)
Download all attachments

2022-03-12 03:15:06

by Stephen Boyd

[permalink] [raw]
Subject: Re: [PATCH v2 0/7] clk: Declare mux tables as const u32[]

Quoting Jonathan Neuschäfer (2022-02-26 04:40:19)
> On Fri, Feb 25, 2022 at 05:00:31PM -0800, Stephen Boyd wrote:
> > Quoting Jonathan Neuschäfer (2022-02-05 02:36:06)
> > > I noticed that the 'table' parameter to clk_register_mux_table is never
> > > used for modifying the table elements, and so it can be declared const.
> > >
> > > In version 2 I'm addressing two warnings in the clk-lpc18xx-cgu driver
> > > that I previously missed.
> >
> > The format of these patches deeply confused my scripts that use git
> > interpret-trailer. I fixed it now, hopefully it keeps working. In the
> > future, please don't add more triple dash '---' sections to the patch.
> > It looks like those extra sections for the changelog messed everything
> > up. Or there's a new bug in interpret-trailers. Either way,
> > interpret-trailers was adding tags after the entire patch contents
> > because I think it looks for the last triple dash instead of the first
> > triple dash. Not sure why it's done that way. I resorted to
> > reconstructing the patch after splitting it with mailinfo.
>
> Hmm, sorry about that.
>
> I've used this format for a while, because it conveniently lets me
> keep my remarks in a git commit (rather than a patch file), until I use
> git format-patch to generate the final patch file.
>
> I'm not very familiar with git interpret-trailers, but git 2.34.1
> doesn't seem to get confused on my patches (or I didn't pass the right
> options to cause it to happen).
>

It's possible it's a git bug. No worries!