GH-386: Add clang-format config and scripts

.clang-format

@@ -0,0 +1,98 @@
+# Clang-format configuration for Zeek. This configuration requires
+# at least clang-format 12.0.1 to format correctly.
+#
+# The easiest way to run this from the command-line is using the
+# python script in auxil/run-clang-format:
+#
+# python3 auxil/run-clang-format/run-clang-format.py --clang-format-executable /path/to/clang-format -r src -i
+
+Language: Cpp
+Standard: c++17
+
+BreakBeforeBraces: Whitesmiths
+
+# BraceWrapping:
+#   AfterCaseLabel: true
+#   AfterClass: false
+#   AfterControlStatement: Always
+#   AfterEnum: false
+#   AfterFunction: true
+#   AfterNamespace: false
+#   AfterStruct: false
+#   AfterUnion: false
+#   AfterExternBlock: false
+#   BeforeCatch: true
+#   BeforeElse: true
+#   BeforeWhile: false
+#   IndentBraces: true
+#   SplitEmptyFunction: false
+#   SplitEmptyRecord: false
+#   SplitEmptyNamespace: false
+
+AccessModifierOffset: -4
+AlignAfterOpenBracket: true
+AlignTrailingComments: false
+AllowShortBlocksOnASingleLine: Empty
+AllowShortEnumsOnASingleLine: true
+AllowShortFunctionsOnASingleLine: Inline
+AllowShortIfStatementsOnASingleLine: false
+AllowShortLambdasOnASingleLine: Empty
+AllowShortLoopsOnASingleLine: false
+AlwaysBreakAfterReturnType: None
+BinPackArguments: true
+BinPackParameters: true
+BreakConstructorInitializers: BeforeColon
+BreakInheritanceList: BeforeColon
+ColumnLimit: 100
+ConstructorInitializerAllOnOneLineOrOnePerLine: false
+FixNamespaceComments: false
+IndentCaseLabels: true
+IndentCaseBlocks: true
+IndentExternBlock: NoIndent
+IndentPPDirectives: None
+IndentWidth: 4
+NamespaceIndentation: None
+PointerAlignment: Left
+SpaceAfterCStyleCast: false
+SpaceAfterLogicalNot: true
+SpaceBeforeAssignmentOperators: true
+SpaceBeforeCpp11BracedList: false
+SpaceBeforeCtorInitializerColon: true
+SpaceBeforeInheritanceColon: true
+SpaceBeforeParens: ControlStatements
+SpaceBeforeRangeBasedForLoopColon: true
+SpaceInEmptyBlock: true
+SpaceInEmptyParentheses: false
+SpacesInAngles: false
+SpacesInConditionalStatement: true
+SpacesInContainerLiterals: false
+SpacesInParentheses: false
+TabWidth: 4
+UseTab: AlignWithSpaces
+
+IncludeBlocks: Regroup
+
+# Include categories go like this:
+# 0: reserved, since this automatically is the primary header for any .cc file
+# 1: zeek-config.h
+# 2: any c-style header
+# 3: any c++-style header
+# 4: any header that starts with "zeek/"
+# 5: everything else, which should catch any of the auto-generated code from the
+#    build directory as well
+#
+# Sections 0-1 and 2-3 get group together in their respective blocks
+IncludeCategories:
+  - Regex: '^"zeek-config\.h"'
+    Priority: 0
+    SortPriority: 1
+  - Regex: '^<[[:print:]]+\.(h|hh)>'
+    Priority: 2
+    SortPriority: 2
+  - Regex: '^<[[:print:]]+>'
+    Priority: 2
+    SortPriority: 3
+  - Regex: '^"zeek/'
+    Priority: 4
+  - Regex: '.*'
+    Priority: 5

.clang-format-ignore

@@ -0,0 +1,17 @@
+# Ignore everything 3rdparty
+src/3rdparty/*
+
+# These are files that are technically sourced from other places but aren't in 3rdparty
+# and shouldn't be reformatted.
+src/ConvertUTF.*
+src/bro_inet_ntop.*
+src/bsd-getopt-long.*
+src/in_cksum.*
+src/nb_dns.*
+src/modp_numtoa.*
+src/patricia.*
+src/strsep.c
+src/setsignal.c
+
+# These files are generated code
+src/DebugCmdInfoConstants.*

.gitmodules

@@ -49,3 +49,6 @@
 [submodule "auxil/zeek-client"]
 	path = auxil/zeek-client
 	url = https://github.com/zeek/zeek-client
+[submodule "auxil/run-clang-format"]
+	path = auxil/run-clang-format
+	url = https://github.com/Sarcasm/run-clang-format

auxil/run-clang-format

@@ -0,0 +1 @@
+Subproject commit 39081c9c42768ab5e8321127a7494ad1647c6a2f

ci/run-clang-format.sh

@@ -0,0 +1,65 @@
+#! /bin/sh
+#
+# Copyright (c) 2020 by the Zeek Project. See LICENSE for details.
+
+base=$(git rev-parse --show-toplevel)
+fix=0
+pre_commit_hook=0
+
+# Directories to run on by default. When changing, adapt .pre-commit-config.yam
+# as well.
+files="src"
+
+error() {
+    test "${pre_commit_hook}" = 0 && echo "$@" >&2 && exit 1
+    exit 0
+}
+
+if [ $# != 0 ]; then
+    case "$1" in
+        --fixit)
+            shift
+            fix=1
+            ;;
+
+        --pre-commit-hook)
+            shift
+            fix=1
+            pre_commit_hook=1
+            ;;
+
+        -*)
+            echo "usage: $(basename $0) [--fixit | --pre-commit-hook] [<files>]"
+            exit 1
+    esac
+fi
+
+test $# != 0 && files="$@"
+
+if [ -z "${CLANG_FORMAT}" ]; then
+    CLANG_FORMAT=$(which clang-format 2>/dev/null)
+fi
+
+if [ -z "${CLANG_FORMAT}" -o ! -x "${CLANG_FORMAT}" ]; then
+    error "Cannot find clang-format. If not in PATH, set CLANG_FORMAT."
+fi
+
+if ! (cd / && ${CLANG_FORMAT} -dump-config | grep -q SpacesInConditionalStatement); then
+    error "${CLANG_FORMAT} does not support SpacesInConditionalStatement. Install custom version and put it into PATH, or point CLANG_FORMAT to it."
+fi
+
+if [ ! -e .clang-format ]; then
+    error "Must execute in top-level directory."
+fi
+
+cmd="${base}/auxil/run-clang-format/run-clang-format.py -r --clang-format-executable ${CLANG_FORMAT} --exclude '*/3rdparty/*' ${files}"
+tmp=/tmp/$(basename $0).$$.tmp
+trap "rm -f ${tmp}" EXIT
+eval "${cmd}" >"${tmp}"
+
+if [ "${fix}" = 1 ]; then
+    test -s "${tmp}" && cat "${tmp}" | git apply -p0
+    true
+else
+    cat "${tmp}"
+fi

src/Func.cc

@@ -55,6 +55,9 @@
 #include "zeek/iosource/PktSrc.h"
 #include "zeek/iosource/PktDumper.h"
 
+// Ignore clang-format's reordering of include files here so that it doesn't
+// break what symbols are available when, which keeps the build from breaking.
+// clang-format off
 #include "zeek.bif.func_h"
 #include "stats.bif.func_h"
 #include "reporter.bif.func_h"
@@ -72,6 +75,7 @@
 #include "supervisor.bif.func_def"
 #include "packet_analysis.bif.func_def"
 #include "CPP-load.bif.func_def"
+// clang-format on
 
 extern	RETSIGTYPE sig_handler(int signo);