--- /sys/src/cmd/tee.c +++ tee.c @@ -9,7 +9,6 @@ FDSTART = 3, }; -int uflag; int aflag; char in[8192]; @@ -32,7 +31,6 @@ break; case 'u': - uflag++; /* uflag is ignored and undocumented; it's a relic from Unix */ break;
ping
Nov 9, 2022 17:17:03 mkf9@riseup.net:
> --- /sys/src/cmd/tee.c
> +++ tee.c
> @@ -9,7 +9,6 @@
> FDSTART = 3,
> };
>
> -int uflag;
> int aflag;
>
> char in[8192];
> @@ -32,7 +31,6 @@
> break;
>
> case 'u':
> - uflag++;
> /* uflag is ignored and undocumented; it's a relic from Unix */
> break;
Applied, thanks. Your original patch looks wrong - git/import could not apply it. Git/export is supposed to produce the right thing.
Sigrid Solveig Haflínudóttir <sigrid@ftrv.se> wrote:
> Your original patch looks wrong - git/import could
> not apply it. Git/export is supposed to produce the
> right thing.
Although the patch isn't what you'd get with git/commit and git/export,
I don't think anything is wrong with it apart from being a bit
unconventional. We should be able to deal with it with our tools.
There are two issues here:
1. The patch relies on standard filename detection behavior that
our new patch implementation doesn't follow. There are plenty
of patches in the wild like this, and I mentioned it in the
`patch(1) filename detection thread` a while back. I have a local
patch patch in my queue that fixes this.
2. git/import has no way to forward -p arguments to patch. git am
has this, and it seems both useful and easy to add to git/import.
With both of those fixed, git/import -p0 would do the trick.
Perhaps the original email should have had a subject like [DIFF] ...
Quoth Michael Forney <mforney@mforney.org>:
> Sigrid Solveig Haflínudóttir <sigrid@ftrv.se> wrote:
> > Your original patch looks wrong - git/import could
> > not apply it. Git/export is supposed to produce the
> > right thing.
>
> Although the patch isn't what you'd get with git/commit and git/export,
> I don't think anything is wrong with it apart from being a bit
> unconventional. We should be able to deal with it with our tools.
>
> There are two issues here:
> 1. The patch relies on standard filename detection behavior that
> our new patch implementation doesn't follow. There are plenty
> of patches in the wild like this, and I mentioned it in the
> `patch(1) filename detection thread` a while back. I have a local
> patch patch in my queue that fixes this.
> 2. git/import has no way to forward -p arguments to patch. git am
> has this, and it seems both useful and easy to add to git/import.
>
> With both of those fixed, git/import -p0 would do the trick.
seems sane to me.
thanks, will use it for next patchs.
Nov 16, 2022 05:18:33 Sigrid Solveig Haflínudóttir <sigrid@ftrv.se>:
> Applied, thanks.
> Your original patch looks wrong - git/import could
> not apply it. Git/export is supposed to produce the
> right thing.