On Mon, Nov 06, 2023 at 10:19:23AM +0100, Martin Kletzander wrote: > On Fri, Nov 03, 2023 at 09:37:19AM -0400, Daniel P. Berrangé wrote: > > Temporarily disable '-Wmissing-include-dirs' becuase the > > libtirpc pkg-confog file has a bogus include dir. The > > headers can fortunately still be found since they are > > in the system include dir. > > > > Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com&gt; > > even though this patch does not apply for me for some reason, no idea why, it is > valid and our CI fails without it, hence > Ah, since the switch to the new ML I cannot apply any patch using git am: warning: quoted CRLF detected Applying: ci: workaround broken opensuse leap 15.5 tirpc Using index info to reconstruct a base tree... error: patch failed: ci/gitlab/builds.yml:332 error: ci/gitlab/builds.yml: patch does not apply error: patch failed: ci/manifest.yml:205

On Mon, Nov 06, 2023 at 10:37:28AM +0100, Martin Kletzander wrote: > On Mon, Nov 06, 2023 at 10:19:23AM +0100, Martin Kletzander wrote: > > On Fri, Nov 03, 2023 at 09:37:19AM -0400, Daniel P. Berrangé wrote: > > > Temporarily disable '-Wmissing-include-dirs' becuase the > > > libtirpc pkg-confog file has a bogus include dir. The > > > headers can fortunately still be found since they are > > > in the system include dir. > > > > > > Signed-off-by: Daniel P. Berrangé <berrange(a)redhat.com&gt; > > > > even though this patch does not apply for me for some reason, no idea why, it is > > valid and our CI fails without it, hence > > Ah, since the switch to the new ML I cannot apply any patch using git > am: > > warning: quoted CRLF detected > Applying: ci: workaround broken opensuse leap 15.5 tirpc > Using index info to reconstruct a base tree... > error: patch failed: ci/gitlab/builds.yml:332 > error: ci/gitlab/builds.yml: patch does not apply > error: patch failed: ci/manifest.yml:205 Yeah, both me and Laine noticed this the other day. Patches look fine in mutt, but if you save them to a file they end up looking all wonky (see attached file for an example) and 'git am' won't like them.

On Mon, Nov 06, 2023 at 09:58:02AM +0000, Daniel P. Berrangé wrote: > On Mon, Nov 06, 2023 at 01:47:20AM -0800, Andrea Bolognani wrote: > > On Mon, Nov 06, 2023 at 10:37:28AM +0100, Martin Kletzander wrote: > > > Ah, since the switch to the new ML I cannot apply any patch using git > > > am: > > > > > > warning: quoted CRLF detected > > > Applying: ci: workaround broken opensuse leap 15.5 tirpc > > > Using index info to reconstruct a base tree... > > > error: patch failed: ci/gitlab/builds.yml:332 > > > error: ci/gitlab/builds.yml: patch does not apply > > > error: patch failed: ci/manifest.yml:205 > > > > Yeah, both me and Laine noticed this the other day. Patches look fine > > in mutt, but if you save them to a file they end up looking all wonky > > (see attached file for an example) and 'git am' won't like them. > > Well that's strange. > > The version I sent was > > Content-Type: text/plain; charset=UTF-8 > Content-Transfer-Encoding: 8bit > > What you have attached is > > Content-Type: text/plain; charset="utf-8" > Content-Transfer-Encoding: base64 > > Meanwhile when downloading from the archives > > Content-Type: text/plain; charset="utf-8" > Content-Transfer-Encoding: quoted-printable > > IOW, something in the mail chain is screwing around with the content. > > Aside from harming git am, this is almost certainly going to doom > DKIM signatures and GPG signatures covering the content. So does that mean that the issue is server-side? I'm pretty sure Laine, Martin and I are all fetching emails using different tools, so that would sound more plausible, but it's weird that things apparently work fine for you?

On Mon, Nov 06, 2023 at 10:26:36AM +0000, Daniel P. Berrangé wrote: > On Mon, Nov 06, 2023 at 02:21:16AM -0800, Andrea Bolognani wrote: >> So does that mean that the issue is server-side? I'm pretty sure >> Laine, Martin and I are all fetching emails using different tools, so >> that would sound more plausible, but it's weird that things >> apparently work fine for you? > > I guess I tested the wrong mail. I can apply patches from Laine, but > I cannot apply from Michal. > > I think the trigger is non-ASCII7 characters (ie accents) in the sender > name (or possibly the Signed-off-by), causing software to re-encode turn > the message into base64 Until this is sorted out, passing --quoted-cr=strip to 'git am' seems to do the trick.

On Mon, Nov 06, 2023 at 12:45:22PM +0100, Michal Prívozník wrote: > On 11/6/23 11:54, Andrea Bolognani wrote: >> On Mon, Nov 06, 2023 at 10:26:36AM +0000, Daniel P. Berrangé wrote: >>> On Mon, Nov 06, 2023 at 02:21:16AM -0800, Andrea Bolognani wrote: >>>> So does that mean that the issue is server-side? I'm pretty sure >>>> Laine, Martin and I are all fetching emails using different tools, so >>>> that would sound more plausible, but it's weird that things >>>> apparently work fine for you? >>> >>> I guess I tested the wrong mail. I can apply patches from Laine, but >>> I cannot apply from Michal. >>> >>> I think the trigger is non-ASCII7 characters (ie accents) in the sender >>> name (or possibly the Signed-off-by), causing software to re-encode turn >>> the message into base64 >> >> Until this is sorted out, passing >> >> --quoted-cr=strip >> >> to 'git am' seems to do the trick. >> > > Alternatively, might as well start sending base64 encoded patches, list > re-encodes them anyway: > > git send-email --transfer-encoding=base64 ... Please don't. base64 encoding is horrible if someone needs to just view mbox contents directly.