Add .clang-format file with default Google style

[phcerdan]

A lot of IDE's, and editors plugins interface well with clang-tools.
This adds the file .clang-format in the top folder (the expected place).
It is populated with the default google style.

It was generated with the command:

clang-format -style=google -dump-config > .clang-format

With

clang-format --version
clang-format version 10.0.0

This should conform with the same style using cpplint.

See #317

---

This change is 

[maxitg]

This is great! But keep in mind, we have exceptions described here. Specifically, we at least need the equivalent of cpplint.py --linelength=120 --filter=-legal/copyright. It does not look like clang-format checks for legal/copyright, but we need lines, and check other comments below.

Reviewed 1 of 1 files at r1.
Reviewable status: all files reviewed, 17 unresolved discussions (waiting on @aokellermann and @phcerdan)

---

.clang-format, line 15 at r1 (raw file):

AllowAllConstructorInitializersOnNextLine: true
AllowAllParametersOfDeclarationOnNextLine: true
AllowShortBlocksOnASingleLine: Never

Should we set this to Empty? @aokellermann?

---

.clang-format, line 17 at r1 (raw file):

AllowShortBlocksOnASingleLine: Never
AllowShortCaseLabelsOnASingleLine: false
AllowShortFunctionsOnASingleLine: All

Should we set this to Empty as well? Not sure, but sounds like it should be consistent with AllowShortBlocksOnASingleLine.

---

.clang-format, line 19 at r1 (raw file):

AllowShortFunctionsOnASingleLine: All
AllowShortLambdasOnASingleLine: All
AllowShortIfStatementsOnASingleLine: WithoutElse

I think we should set this to Always because it's dangerous otherwise. We don't want code like this:

if (a) return;
else
  return;

because it could easily become before anybody notices:

if (a) return;
else
  a = false;
  return;

---

.clang-format, line 21 at r1 (raw file):

AllowShortIfStatementsOnASingleLine: WithoutElse
AllowShortLoopsOnASingleLine: true
AlwaysBreakAfterDefinitionReturnType: None

The docs say it's deprecated, so we should probably remove it.

---

.clang-format, line 26 at r1 (raw file):

AlwaysBreakTemplateDeclarations: Yes
BinPackArguments: true
BinPackParameters: true

Should we set these two to false to make arguments more obvious? Not sure.

---

.clang-format, line 27 at r1 (raw file):

BinPackArguments: true
BinPackParameters: true
BraceWrapping:

This is ignored because BreakBeforeBraces is not set to BS_Custom

---

.clang-format, line 53 at r1 (raw file):

BreakAfterJavaFieldAnnotations: false
BreakStringLiterals: true
ColumnLimit:     80

This should be 120.

---

.clang-format, line 54 at r1 (raw file):

BreakStringLiterals: true
ColumnLimit:     80
CommentPragmas:  '^ IWYU pragma:'

If I understand how this works, I think we should set this to '^ NO LINT' to be consistent with cpplint.py.

---

.clang-format, line 61 at r1 (raw file):

Cpp11BracedListStyle: true
DeriveLineEnding: true
DerivePointerAlignment: true

I think we should set this to false and make the alignment consistent everywhere.

---

.clang-format, line 65 at r1 (raw file):

ExperimentalAutoDetectBinPacking: false
FixNamespaceComments: true
ForEachMacros:

I don't think we need this since we are not using any of those.

---

.clang-format, line 91 at r1 (raw file):

IndentWrappedFunctionNames: false
JavaScriptQuotes: Leave
JavaScriptWrapImports: true

I think we should remove the JavaScript* configs since we don't plan to have any. We can decide on the style if/when we do.

---

.clang-format, line 100 at r1 (raw file):

ObjCBlockIndentWidth: 2
ObjCSpaceAfterProperty: false
ObjCSpaceBeforeProtocolList: true

Same for ObjC*.

---

.clang-format, line 122 at r1 (raw file):

    CanonicalDelimiter: ''
    BasedOnStyle:    google
  - Language:        TextProto

Same here.

---

.clang-format, line 155 at r1 (raw file):

SpacesInAngles:  false
SpacesInConditionalStatement: false
SpacesInContainerLiterals: true

I don't think this applies to C++, but if it does, it should be false.

---

.clang-format, line 160 at r1 (raw file):

SpacesInSquareBrackets: false
SpaceBeforeSquareBrackets: false
Standard:        Auto

Should we say c++17 explicitly to avoid any potential incorrect detection?

---

.clang-format, line 161 at r1 (raw file):

SpaceBeforeSquareBrackets: false
Standard:        Auto
StatementMacros:

I don't think we need any of these either.

---

.clang-format, line 164 at r1 (raw file):

  - Q_UNUSED
  - QT_REQUIRE_VERSION
TabWidth:        8

We don't use tabs, so we should either remove it, or at least set it to 2.

[phcerdan]

Reviewable status: all files reviewed, 18 unresolved discussions (waiting on @aokellermann and @phcerdan)

a discussion (no related file):
All those options are used. It would be somehow cleaner to only write the changes that differ from those.
And uncomment BasedOnStyle: Google.

---

[maxitg]

Reviewed 1 of 1 files at r2.
Reviewable status: all files reviewed, 6 unresolved discussions (waiting on @aokellermann and @phcerdan)

a discussion (no related file):

Previously, phcerdan (Pablo Hernandez-Cerdan) wrote…

All those options are used. It would be somehow cleaner to only write the changes that differ from those.
And uncomment BasedOnStyle: Google.

I agree, we should do that.

---

[phcerdan]

Reviewable status: 0 of 1 files reviewed, 6 unresolved discussions (waiting on @aokellermann, @maxitg, and @phcerdan)

a discussion (no related file):

Previously, maxitg (Max Piskunov) wrote…

I agree, we should do that.

When you completely decide about option, we can commit and apply the clang-format to all files. Right now check an example on these options applied to Expressions.cpp:

---

---

.clang-format, line 53 at r1 (raw file):

Previously, maxitg (Max Piskunov) wrote…

This should be 120.

Done.

---

.clang-format, line 54 at r1 (raw file):

Previously, maxitg (Max Piskunov) wrote…

If I understand how this works, I think we should set this to '^ NO LINT' to be consistent with cpplint.py.

It seems to be NOLINT:.*