From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: mx.groups.io; dkim=missing; spf=pass (domain: insyde.com, ip: 210.71.195.41, mailfrom: kevin.davis@insyde.com) Received: from out01.hibox.biz (out01.hibox.biz [210.71.195.41]) by groups.io with SMTP; Thu, 04 Apr 2019 10:17:58 -0700 X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: =?us-ascii?q?A2ADAAATOaZc/ww0GKxlGQEBAQEBAQE?= =?us-ascii?q?BAQEBAQcBAQEBAQGBUQQBAQEBAQsBgQ4vAYE6BX4nhA6IHF+MJCWDXZRmFIF?= =?us-ascii?q?nECMIAYFLgnUCIIVPNAkNAQEDAQEJAQMCbRwMhUoBAQICASMrKwUHDAMNBAQ?= =?us-ascii?q?BAQEgBwMCMxUJCAcSHQSDAQGBbQkOrjWBLxoChBUBhgGBMAGBWosvP4ERJww?= =?us-ascii?q?Tgh4uPoJhBBiBMBc2glQxgiYDikyHVYEchVGLemIHAogAi3oagmKJAQOIaYt?= =?us-ascii?q?PhhyCHYhEgxuBTziBVk0jegGCQQk1gVgXhjCCMFGFCiYwAYEAAQYBAQEVCBO?= =?us-ascii?q?PLQEB?= X-IronPort-AV: E=Sophos;i="5.60,309,1549900800"; d="scan'208,217";a="34351805" Received: from unknown (HELO hb3-BKT202.hibox.biz) ([172.24.52.12]) by out01.hibox.biz with ESMTP; 05 Apr 2019 01:17:53 +0800 Received: from unknown (HELO hb3-BKT101.hibox.biz) ([172.24.51.11]) by hb3-BKT202.hibox.biz with ESMTP; 05 Apr 2019 01:17:52 +0800 Received: from unknown (HELO hb3-IN02.hibox.biz) ([172.24.12.12]) by hb3-BKT101.hibox.biz with ESMTP; 05 Apr 2019 01:17:53 +0800 X-Remote-IP: 172.24.131.45 X-Remote-Host: No hostname X-SBRS: None X-MID: 21908067 X-Auth-ID: kevin.davis@insyde.com X-EnvelopeFrom: kevin.davis@insyde.com Received: from unknown (HELO [IPv6:2600:380:a118:b5d2:498e:9d44:369d:6e34]) ([172.24.131.45]) by hb3-IN02.hibox.biz with ESMTP/TLS/AES256-SHA; 05 Apr 2019 01:17:50 +0800 Mime-Version: 1.0 (1.0) Subject: Re: [edk2-devel] [Patch 0/4] Normalize line endings to CRLF in ARM packages From: "Kevin@Insyde" X-Mailer: iPhone Mail (16E227) In-Reply-To: Date: Thu, 4 Apr 2019 13:17:46 -0400 Cc: Laszlo Ersek , Leif Lindholm , Ard Biesheuvel , Andrew Fish Message-Id: To: devel@edk2.groups.io, michael.d.kinney@intel.com Content-Type: multipart/alternative; boundary=Apple-Mail-F24DC1D7-3ECE-4F0B-B45F-E5289CB9060B Content-Transfer-Encoding: 7bit --Apple-Mail-F24DC1D7-3ECE-4F0B-B45F-E5289CB9060B Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: quoted-printable We would prefer separate patches. Kevin D Davis Insyde Software > On Apr 4, 2019, at 1:00 PM, Michael D Kinney wrote: >=20 > Leif, >=20 > There were a few files when the license change patches > were created that failed PatchCheck.py. >=20 > I am trying to resolve those before the license change > so the license change patches do not have any PatchCheck.py > failures. >=20 > If I do these as part of the license change, then the > change will look larger than just the license change in > these files. >=20 > I prefer to have a separate patch to resolve the line=20 > endings. >=20 > There may be other files in edk2 repo that have mixed > line endings in lines not associated with the license > change itself. We can consider cleaning those up in > a separate series. >=20 > The other option is to abandon these patches to normalize > line endings and allow the license change patches to not > pass PatchCheck.py. >=20 > Which do you prefer? >=20 > Thanks, >=20 > Mike >=20 >=20 >=20 >> -----Original Message----- >> From: Laszlo Ersek [mailto:lersek@redhat.com] >> Sent: Thursday, April 4, 2019 3:39 AM >> To: Leif Lindholm ; Kinney, >> Michael D >> Cc: devel@edk2.groups.io; Ard Biesheuvel >> ; Andrew Fish >> >> Subject: Re: [Patch 0/4] Normalize line endings to CRLF >> in ARM packages >>=20 >>> On 04/04/19 05:54, Leif Lindholm wrote: >>> Hi Mike, >>>=20 >>> This looks fine. But I did have one thought: >>> If we're just about to touch pretty much every file in >> the tree anyway >>> - can we consider doing the full conversion to native >> line endings at >>> the same time? >>>=20 >>> (This doesn't help the 'git blame' problem, it just >> reminded me.) >>=20 >> The git-blame issue that such a conversion introduces is >> not terrible. >> Let's say you run >>=20 >> git blame -- ArmPkg/Library/GccLto/liblto-arm.s >>=20 >> and it gives you a commit for a specific line you care >> about. Let's call >> that commit C1. >>=20 >> Then you do >>=20 >> git show --color C1 | less >>=20 >> and expect to see the commit message & patch that >> introduced the *logic* >> on that line. But, you only see a "useless" CRLF >> conversion, in that >> commit. So what do you do? Simple, repeat the process as >> follows: >>=20 >> git blame C1^ -- ArmPkg/Library/GccLto/liblto-arm.s >>=20 >> This will assign blame on the lines of the file as the >> file was right >> before C1 changed the line terminators. >>=20 >>=20 >> The problem that I think *is* worse is that, jumping >> about in the git >> history (with git checkout / git bisect), around the >> large conversion >> commit, will expose the user in a brief time to files >> that are both LF >> and CRLF encoded, internally. Therefore, whatever their >> local >> auto-conversion-on-checkout settings are, some commits >> in such a work >> session will not match those settings. >>=20 >> For example, assuming you are on Windows, and set up >> LF-to-CRLF-on-checkout (as you should), then git- >> checkout a commit from >> before the "LF only" conversion commit, you might see >> double CRs. I >> guess. I never tested it. >>=20 >> Thanks >> Laszlo >>=20 >>=20 >>=20 >>>=20 >>> / >>> Leif >>>=20 >>> On Wed, Apr 03, 2019 at 03:00:10PM -0700, Kinney, >> Michael D wrote: >>>> https://bugzilla.tianocore.org/show_bug.cgi?id=3D1659 >>>>=20 >>>> Normalize line endings to use CRLF to pass >> PatchCheck.py >>>>=20 >>>> Cc: Leif Lindholm >>>> Cc: Ard Biesheuvel >>>> Cc: Laszlo Ersek >>>> Contributed-under: TianoCore Contribution Agreement >> 1.1 >>>> Signed-off-by: Michael D Kinney >> >>>>=20 >>>> Kinney, Michael D (4): >>>> ArmPkg/ArmScmiDxe: Remove non-ASCII character in >> comment >>>> ArmPkg: Normalize line endings to CRLF >>>> ArmVirtPkg: Normalize line endings to CRLF >>>> ArmPlatformPkg: Normalize line endings to CRLF >>>>=20 >>>> .../ArmScmiPerformanceProtocolPrivate.h | 2 >> +- >>>> ArmPkg/Library/ArmSmcLibNull/ArmSmcLibNull.c | 44 >> +- >>>> .../ArmSmcPsciResetSystemLib/AArch64/Reset.S | 60 >> +- >>>> .../AArch64/Reset.asm | 70 >> +- >>>> .../ArmSmcPsciResetSystemLib/Arm/Reset.S | 58 >> +- >>>> .../ArmSmcPsciResetSystemLib/Arm/Reset.asm | 68 >> +- >>>> ArmPkg/Library/CompilerIntrinsicsLib/memset.c | 124 >> ++-- >>>> ArmPkg/Library/GccLto/liblto-aarch64.s | 54 >> +- >>>> ArmPkg/Library/GccLto/liblto-arm.s | 122 >> ++-- >>>> ArmPlatformPkg/Scripts/Ds5/profile.py | 668 >> +++++++++--------- >>>> ArmVirtPkg/Include/Platform/Hidden.h | 56 >> +- >>>> 11 files changed, 663 insertions(+), 663 deletions(- >> ) >>>>=20 >>>> -- >>>> 2.21.0.windows.1 >>>>=20 >=20 >=20 >=20 >=20 --Apple-Mail-F24DC1D7-3ECE-4F0B-B45F-E5289CB9060B Content-Type: text/html; charset=utf-8 Content-Transfer-Encoding: quoted-printable We would prefer separate patches.
<= br>
=

Kevin D Davis
Insyde Software


On Apr 4, 2019, at 1:00 PM, M= ichael D Kinney <michael.d= .kinney@intel.com> wrote:

Leif,

There were a few= files when the license change patches
were created that fa= iled PatchCheck.py.

I am trying to resolve= those before the license change
so the license change patc= hes do not have any PatchCheck.py
failures.

If I do these as part of the license change, then the
change will look larger than just the license change in=
these files.

I prefer to have a = separate patch to resolve the line
endings.

There may be other files in edk2 repo that have mixed
line endings in lines not associated with the license<= br>change itself.  We can consider cleaning those up ina separate series.

The other optio= n is to abandon these patches to normalize
line endings and= allow the license change patches to not
pass PatchCheck.py= .

Which do you prefer?

Thanks,

Mike



-----Original Message-----
From: Laszlo Ersek [mailto:l= ersek@redhat.com]
Sent: Thursday, April 4, 2019 3:39 AM
To: Leif Lindholm <leif.lindholm@linaro.org>; Kinney,
Michael D <michael.d.kinney@intel.com>
Cc: devel@edk2.groups.io; Ard Biesheuvel
<ard= .biesheuvel@linaro.org>; Andrew Fish
<afish@apple.= com>
Subject:= Re: [Patch 0/4] Normalize line endings to CRLF
in ARM packages

<= span>On 04/04/19 05:54, Leif Lindholm wrote:
Hi Mike,

This looks fine. But I did have one thought:=
If we're just about to touch pretty much every file in<= br>
the tree anywa= y
- can we consider doing the full conversion to native
line endings at
the same time?

(This doesn't help = the 'git blame' problem, it just
reminded me.)

The git-blame issue that such a conversion introduces is
not terrible.
=
Let's say you run

 git blame -- ArmPkg/Library/GccLto/liblto-arm.s

and it gives you a commit for a specific line yo= u care
about. Let's = call
that commit C1.=

Then you do

 git show --color C1 | less

<= span>and expect to see the commit message & patch that
introduced the *logic*
on that line. But, you only see a= "useless" CRLF
conv= ersion, in that
comm= it. So what do you do? Simple, repeat the process as
follows:

 git blame C1^ -- ArmPkg/Library/GccLto/liblto-arm.s

This will assign blame on the lines of the file as t= he
file was right
before C1 changed the = line terminators.


The problem that I think *is* worse is t= hat, jumping
about i= n the git
history (w= ith git checkout / git bisect), around the
large conversion
commit, will expose the user in a brief time to files
that are both LF
and CRLF encoded, intern= ally. Therefore, whatever their
local
au= to-conversion-on-checkout settings are, some commits
in such a work
session will not match those settings.

For example, assuming you are on Windows, and s= et up
LF-to-CRLF-on-= checkout (as you should), then git-
checkout a commit from
before the "LF only" conversion commit, you might see
double CRs. I<= br>
guess. I never tested it.

Thanks
Laszlo



/<= /span>
   Leif

On Wed, Apr 03, 2019 at 03:00:10PM -0700, Kinney,
Michael D wrote:
https://bugzilla.tianocore.org/show_bug.cgi?id=3D1659

Normalize line endings to use CR= LF to pass
PatchCheck.py

<= /blockquote>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Ard Biesheuvel &l= t;ard.biesheuvel@linaro.org>
Cc: Laszlo = Ersek <lersek@redhat.com>
Contributed-under: = TianoCore Contribution Agreement
1.1
= Signed-off-by: Michael D Kinney
<michael.d.kinney@intel.com>
<= span>
Kinney, Mic= hael D (4):
&nbs= p;ArmPkg/ArmScmiDxe: Remove non-ASCII character in
<= /blockquote>
comment
=
 ArmPkg: Normalize line endings to CRLF<= br>
 ArmVirtPkg: Norma= lize line endings to CRLF
=
 ArmPlatformPkg: Normalize line endings to CRLF

.../ArmScmiPerformanceProtocolPrivate.h &nbs= p;     |   2
+-
ArmPkg/Library/ArmSmcLibNull/ArmSmcLibNull.c  | &nbs= p;44
+-
.../ArmSmcPsciResetSystemLib= /AArch64/Reset.S  |  60
+-
= .../AArch64/Reset.asm         &nbs= p;            &= nbsp;  |  70
+-
.../A= rmSmcPsciResetSystemLib/Arm/Reset.S      |  5= 8
+-
.../ArmSmcPsciResetSystemLib/Ar= m/Reset.asm    |  68
+-
ArmPkg/Library/CompilerIntrinsicsLib/memset.c | 124
++--
ArmPkg/Library/GccLto/liblto-aarch64.s  &= nbsp;     |  54
+-
ArmPkg/Library/GccLto/liblto-arm.s     &nb= sp;      | 122
++--
ArmPlatformPkg/Scripts/Ds5/profile.py     = ;    | 668
+++++++++---------
ArmVirtPkg/Include/Platform/Hidden.h     =      |  56
+-
11 files changed, 663 insertions(+), 663 deletions(-
)=

--
2.21= .0.windows.1





--Apple-Mail-F24DC1D7-3ECE-4F0B-B45F-E5289CB9060B--