Added a check for double whitespaces

dbms/src/AggregateFunctions/QuantileExact.h

@@ -176,7 +176,7 @@ struct QuantileExactExclusive : public QuantileExact<Value>
     }
 };
 
-/// QuantileExactInclusive is equivalent to Excel PERCENTILE and PERCENTILE.INC, R-7,  SciPy-(1,1)
+/// QuantileExactInclusive is equivalent to Excel PERCENTILE and PERCENTILE.INC, R-7, SciPy-(1,1)
 template <typename Value>
 struct QuantileExactInclusive : public QuantileExact<Value>
 {

dbms/src/DataTypes/DataTypeEnum.cpp

@@ -234,7 +234,7 @@ void DataTypeEnum<Type>::deserializeBinaryBulk(
 }
 
 template <typename Type>
-void DataTypeEnum<Type>::serializeProtobuf(const IColumn & column, size_t row_num,  ProtobufWriter & protobuf, size_t & value_index) const
+void DataTypeEnum<Type>::serializeProtobuf(const IColumn & column, size_t row_num, ProtobufWriter & protobuf, size_t & value_index) const
 {
     if (value_index)
         return;

dbms/src/Functions/GeoUtils.cpp

@@ -18,7 +18,7 @@ const UInt8 geohash_base32_decode_lookup_table[256] = {
     0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
     0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
     0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
-    0,   1,     2,    3,    4,    5,    6,    7,    8,    9,    0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
+    0,    1,    2,    3,    4,    5,    6,    7,    8,    9,    0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
     0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
     0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF, 0xFF,
     0xFF, 0xFF, 10,   11,   12,   13,   14,   15,   16,   0xFF, 17,   18,   0xFF, 19,   20,   0xFF,

dbms/src/Processors/Formats/Impl/ProtobufRowInputFormat.cpp

@@ -12,7 +12,7 @@ namespace DB
 
 ProtobufRowInputFormat::ProtobufRowInputFormat(ReadBuffer & in_, const Block & header_, Params params_, const FormatSchemaInfo & info_)
     : IRowInputFormat(header_, in_, params_)
-    ,  data_types(header_.getDataTypes())
+    , data_types(header_.getDataTypes())
     , reader(in, ProtobufSchemas::instance().getMessageTypeForFormatSchema(info_), header_.getNames())
 {
 }

dbms/src/Storages/MergeTree/MergeTreeRangeReader.cpp

@@ -108,7 +108,7 @@ size_t MergeTreeRangeReader::DelayedStream::finalize(Block & block)
 
 
 MergeTreeRangeReader::Stream::Stream(
-        size_t from_mark, size_t to_mark,  MergeTreeReader * merge_tree_reader_)
+        size_t from_mark, size_t to_mark, MergeTreeReader * merge_tree_reader_)
         : current_mark(from_mark), offset_after_current_mark(0)
         , last_mark(to_mark)
         , merge_tree_reader(merge_tree_reader_)

dbms/src/Storages/MergeTree/MergeTreeRangeReader.h

@@ -41,7 +41,7 @@ public:
     {
     public:
         DelayedStream() = default;
-        DelayedStream(size_t from_mark,  MergeTreeReader * merge_tree_reader);
+        DelayedStream(size_t from_mark, MergeTreeReader * merge_tree_reader);
 
         /// Read @num_rows rows from @from_mark starting from @offset row
         /// Returns the number of rows added to block.

dbms/src/Storages/System/StorageSystemNumbers.cpp

@@ -146,7 +146,7 @@ BlockInputStreams StorageSystemNumbers::read(
         res[i] = std::make_shared<NumbersBlockInputStream>(max_block_size, offset + i * max_block_size, num_streams * max_block_size);
 
         if (limit)  /// This formula is how to split 'limit' elements to 'num_streams' chunks almost uniformly.
-            res[i] = std::make_shared<LimitBlockInputStream>(res[i], *limit * (i + 1) / num_streams - *limit * i / num_streams, 0,  false, true);
+            res[i] = std::make_shared<LimitBlockInputStream>(res[i], *limit * (i + 1) / num_streams - *limit * i / num_streams, 0, false, true);
     }
 
     return res;

utils/check-style/check-style

@@ -28,3 +28,6 @@ find $ROOT_PATH/dbms -name '*.h' -or -name '*.cpp' |
 
 # Broken symlinks
 find -L $ROOT_PATH -type l | grep -v contrib && echo "^ Broken symlinks found"
+
+# Double whitespaces
+find $ROOT_PATH/dbms -name '*.h' -or -name '*.cpp' | while read i; do $ROOT_PATH/utils/check-style/double-whitespaces.pl < $i || echo -e "^ File $i contains double whitespaces\n"; done

utils/check-style/double-whitespaces.pl

@@ -0,0 +1,33 @@
+#!/usr/bin/perl
+
+use strict;
+
+# Find double whitespace such as "a,  b, c" that looks very ugly and annoying.
+# But skip double whitespaces if they are used as an alignment - by comparing to surrounding lines.
+
+my @array;
+
+while (<>)
+{
+    push @array, $_;
+}
+
+my $ret = 0;
+
+for (my $i = 1; $i < $#array; ++$i)
+{
+    if ($array[$i] =~ ',( {2,3})[^ /]')
+    {
+        # https://stackoverflow.com/questions/87380/how-can-i-find-the-location-of-a-regex-match-in-perl
+
+        if ((substr($array[$i - 1], $+[1] - 1, 2) !~ /^[ -][^ ]$/) # whitespaces are not part of alignment
+         && (substr($array[$i + 1], $+[1] - 1, 2) !~ /^[ -][^ ]$/)
+         && $array[$i] !~ /(-?\d+\w*,\s+){3,}/) # this is not a number table like { 10, -1,  2 }
+        {
+            print(($i + 1) . ":" . $array[$i]);
+            $ret = 1;
+        }
+    }
+}
+
+exit $ret;