Re: [PATCH] Replace HTTP links with HTTPS ones: documentation

From: Alexander A. Klimov
Date: Mon May 25 2020 - 14:01:52 EST




Am 25.05.20 um 18:28 schrieb Jonathan Corbet:
On Wed, 20 May 2020 22:00:37 +0200
"Alexander A. Klimov" <grandmaster@xxxxxxxxxxxx> wrote:

Rationale: Reduces attack surface on kernel devs for MITM.

Deterministic algorithm:
For each file:
For each line:
If doesn't contain `\bxmlns\b`:
For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
If both the HTTP and HTTPS versions
return 200 OK and serve the same content:
Replace HTTP with HTTPS.
---
Documentation/COPYING-logo | 2 +-
Documentation/admin-guide/LSM/tomoyo.rst | 16 ++++++++--------
.../admin-guide/acpi/initrd_table_override.rst | 2 +-
Documentation/admin-guide/bcache.rst | 4 ++--
Documentation/admin-guide/devices.rst | 2 +-
Documentation/admin-guide/initrd.rst | 2 +-
Documentation/admin-guide/md.rst | 2 +-
Documentation/admin-guide/mono.rst | 4 ++--
Documentation/admin-guide/reporting-bugs.rst | 2 +-
Documentation/admin-guide/unicode.rst | 4 ++--
Documentation/conf.py | 2 +-
Documentation/dev-tools/gdb-kernel-debugging.rst | 2 +-
Documentation/doc-guide/parse-headers.rst | 2 +-
.../driver-api/acpi/linuxized-acpica.rst | 6 +++---
Documentation/driver-api/usb/bulk-streams.rst | 4 ++--
.../driver-api/usb/writing_musb_glue_layer.rst | 6 +++---
Documentation/filesystems/path-lookup.txt | 2 +-
Documentation/filesystems/seq_file.txt | 4 ++--
Documentation/misc-devices/c2port.txt | 6 +++---
Documentation/process/3.Early-stage.rst | 2 +-
Documentation/process/7.AdvancedTopics.rst | 8 ++++----
Documentation/process/8.Conclusion.rst | 14 +++++++-------
Documentation/process/adding-syscalls.rst | 4 ++--
Documentation/process/applying-patches.rst | 4 ++--
.../process/volatile-considered-harmful.rst | 4 ++--
Documentation/rbtree.txt | 4 ++--
Documentation/security/SCTP.rst | 2 +-
Documentation/sphinx/kfigure.py | 6 +++---
Documentation/static-keys.txt | 2 +-
Documentation/trace/events-msr.rst | 2 +-
Documentation/trace/mmiotrace.rst | 2 +-
Documentation/vm/ksm.rst | 2 +-
Documentation/xz.txt | 6 +++---
scripts/kernel-doc | 2 +-
34 files changed, 69 insertions(+), 69 deletions(-)

OK, so this is still pretty large; I had asked you to narrow things
further. And the rationale is still pretty thin. And I would *really*
rather see the docs updated in a more thoughtful way that considers the
value of the links rather than just the protocol used.
OK, I'll keep that size rating in mind.

Also I'll try to reason a bit verbosely. (Not my skill #1, sorry.)

I have already one more enhancement idea, but I'd like not to have > 1 open kernel construction areas of mine in parallel (to reduce merge conflicts).


But this does not seem to do harm at this point, so I have gone ahead and
applied it.
What does this mean? Is there a kernel.org repo/branch with my patch?


If you proceed with this work, I'd encourage you to be narrow in your
focus, be sure to copy the right people, and to justify the changes well,
or you may well run into pushback.
I try my best while CCing, but some mailing lists seem to have (silently!) rejected my patch (see discussion on a previous version of my patch).

Does it matter whether I just feed the output of ...

perl scripts/get_maintainer.pl --separator , --norolestats 0001-Replace-HTTP-links-with-HTTPS-ones-documentation.patch

... into Git send-email To: prompt or actually CC: (not To:) mailing lists?


Thanks,

jon