Files
terminal/src/buffer/out/CharRow.cpp
Austin Lamb 539a5dc0af Greatly reduce allocations in the conhost/OpenConsole startup path (#8489)
I was looking at conhost/OpenConsole and noticed it was being pretty
inefficient with allocations due to some usages of std::deque and
std::vector that didn't need to be done quite that way.

So this uses std::vector for the TextBuffer's storage of ROW objects,
which allows one allocation to contiguously reserve space for all the
ROWs - on Desktop this is 9001 ROW objects which means it saves 9000
allocations that the std::deque would have done.  Plus it has the
benefit of increasing locality of the ROW objects since deque is going
to chase pointers more often with its data structure.

Then, within each ROW there are CharRow and ATTR_ROW objects that use
std::vector today.  This changes them to use Boost's small_vector, which
is a variation of vector that allows for the so-called "small string
optimization."  Since we know the typical size of these vectors, we can
pre-reserve the right number of elements directly in the
CharRow/ATTR_ROW instances, avoiding any heap allocations at all for
constructing these objects.

There are a ton of variations on this "small_vector" concept out there
in the world - this one in Boost, LLVM has one called SmallVector,
Electronic Arts' STL has a small_vector, Facebook's folly library has
one...there are a silly number of these out there.  But Boost seems like
it's by far the easiest to consume in terms of integration into this
repo, the CI/CD pipeline, licensing, and stuff like that, so I went with
the boost version.

In terms of numbers, I measured the startup path of OpenConsole.exe on
my dev box for Release x64 configuration.  My box is an i7-6700k @ 4
Ghz, with 32 GB RAM, not that I think machine config matters much here:

|        | Allocation count    | Allocated bytes    | CPU usage (ms) |
| ------ | ------------------- | ------------------ | -------------- |
| Before | 29,461              | 4,984,640          | 103            |
| After  | 2,459 (-91%)        | 4,853,931 (-2.6%)  | 96 (-7%)       |

Along the way, I also fixed a dynamic initializer I happened to spot in
the registry code, and updated some docs.

## Validation Steps Performed
- Ran "runut", "runft" and "runuia" locally and confirmed results are
  the same as the main branch
- Profiled the before/after numbers in the Visual Studio profiler, for
  the numbers shown in the table

Co-authored-by: Austin Lamb <austinl@microsoft.com>
2020-12-16 10:40:30 -08:00

331 lines
8.7 KiB
C++

// Copyright (c) Microsoft Corporation.
// Licensed under the MIT license.
#include "precomp.h"
#include "CharRow.hpp"
#include "unicode.hpp"
#include "Row.hpp"
// Routine Description:
// - constructor
// Arguments:
// - rowWidth - the size (in wchar_t) of the char and attribute rows
// - pParent - the parent ROW
// Return Value:
// - instantiated object
// Note: will through if unable to allocate char/attribute buffers
#pragma warning(push)
#pragma warning(disable : 26447) // small_vector's constructor says it can throw but it should not given how we use it. This suppresses this error for the AuditMode build.
CharRow::CharRow(size_t rowWidth, ROW* const pParent) noexcept :
_wrapForced{ false },
_doubleBytePadded{ false },
_data(rowWidth, value_type()),
_pParent{ FAIL_FAST_IF_NULL(pParent) }
{
}
#pragma warning(pop)
// Routine Description:
// - Sets the wrap status for the current row
// Arguments:
// - wrapForced - True if the row ran out of space and we forced to wrap to the next row. False otherwise.
// Return Value:
// - <none>
void CharRow::SetWrapForced(const bool wrapForced) noexcept
{
_wrapForced = wrapForced;
}
// Routine Description:
// - Gets the wrap status for the current row
// Arguments:
// - <none>
// Return Value:
// - True if the row ran out of space and we were forced to wrap to the next row. False otherwise.
bool CharRow::WasWrapForced() const noexcept
{
return _wrapForced;
}
// Routine Description:
// - Sets the double byte padding for the current row
// Arguments:
// - fWrapWasForced - True if the row ran out of space for a double byte character and we padded out the row. False otherwise.
// Return Value:
// - <none>
void CharRow::SetDoubleBytePadded(const bool doubleBytePadded) noexcept
{
_doubleBytePadded = doubleBytePadded;
}
// Routine Description:
// - Gets the double byte padding status for the current row.
// Arguments:
// - <none>
// Return Value:
// - True if the row didn't have space for a double byte character and we were padded out the row. False otherwise.
bool CharRow::WasDoubleBytePadded() const noexcept
{
return _doubleBytePadded;
}
// Routine Description:
// - gets the size of the row, in glyph cells
// Arguments:
// - <none>
// Return Value:
// - the size of the row
size_t CharRow::size() const noexcept
{
return _data.size();
}
// Routine Description:
// - Sets all properties of the CharRowBase to default values
// Arguments:
// - sRowWidth - The width of the row.
// Return Value:
// - <none>
void CharRow::Reset() noexcept
{
for (auto& cell : _data)
{
cell.Reset();
}
_wrapForced = false;
_doubleBytePadded = false;
}
// Routine Description:
// - resizes the width of the CharRowBase
// Arguments:
// - newSize - the new width of the character and attributes rows
// Return Value:
// - S_OK on success, otherwise relevant error code
[[nodiscard]] HRESULT CharRow::Resize(const size_t newSize) noexcept
{
try
{
const value_type insertVals;
_data.resize(newSize, insertVals);
}
CATCH_RETURN();
return S_OK;
}
typename CharRow::iterator CharRow::begin() noexcept
{
return _data.begin();
}
typename CharRow::const_iterator CharRow::cbegin() const noexcept
{
return _data.cbegin();
}
typename CharRow::iterator CharRow::end() noexcept
{
return _data.end();
}
typename CharRow::const_iterator CharRow::cend() const noexcept
{
return _data.cend();
}
// Routine Description:
// - Inspects the current internal string to find the left edge of it
// Arguments:
// - <none>
// Return Value:
// - The calculated left boundary of the internal string.
size_t CharRow::MeasureLeft() const noexcept
{
const_iterator it = _data.cbegin();
while (it != _data.cend() && it->IsSpace())
{
++it;
}
return it - _data.cbegin();
}
// Routine Description:
// - Inspects the current internal string to find the right edge of it
// Arguments:
// - <none>
// Return Value:
// - The calculated right boundary of the internal string.
size_t CharRow::MeasureRight() const
{
const_reverse_iterator it = _data.crbegin();
while (it != _data.crend() && it->IsSpace())
{
++it;
}
return _data.crend() - it;
}
void CharRow::ClearCell(const size_t column)
{
_data.at(column).Reset();
}
// Routine Description:
// - Tells you whether or not this row contains any valid text.
// Arguments:
// - <none>
// Return Value:
// - True if there is valid text in this row. False otherwise.
bool CharRow::ContainsText() const noexcept
{
for (const value_type& cell : _data)
{
if (!cell.IsSpace())
{
return true;
}
}
return false;
}
// Routine Description:
// - gets the attribute at the specified column
// Arguments:
// - column - the column to get the attribute for
// Return Value:
// - the attribute
// Note: will throw exception if column is out of bounds
const DbcsAttribute& CharRow::DbcsAttrAt(const size_t column) const
{
return _data.at(column).DbcsAttr();
}
// Routine Description:
// - gets the attribute at the specified column
// Arguments:
// - column - the column to get the attribute for
// Return Value:
// - the attribute
// Note: will throw exception if column is out of bounds
DbcsAttribute& CharRow::DbcsAttrAt(const size_t column)
{
return _data.at(column).DbcsAttr();
}
// Routine Description:
// - resets text data at column
// Arguments:
// - column - column index to clear text data from
// Return Value:
// - <none>
// Note: will throw exception if column is out of bounds
void CharRow::ClearGlyph(const size_t column)
{
_data.at(column).EraseChars();
}
// Routine Description:
// - returns text data at column as a const reference.
// Arguments:
// - column - column to get text data for
// Return Value:
// - text data at column
// - Note: will throw exception if column is out of bounds
const CharRow::reference CharRow::GlyphAt(const size_t column) const
{
THROW_HR_IF(E_INVALIDARG, column >= _data.size());
return { const_cast<CharRow&>(*this), column };
}
// Routine Description:
// - returns text data at column as a reference.
// Arguments:
// - column - column to get text data for
// Return Value:
// - text data at column
// - Note: will throw exception if column is out of bounds
CharRow::reference CharRow::GlyphAt(const size_t column)
{
THROW_HR_IF(E_INVALIDARG, column >= _data.size());
return { *this, column };
}
std::wstring CharRow::GetText() const
{
std::wstring wstr;
wstr.reserve(_data.size());
for (size_t i = 0; i < _data.size(); ++i)
{
const auto glyph = GlyphAt(i);
if (!DbcsAttrAt(i).IsTrailing())
{
for (const auto wch : glyph)
{
wstr.push_back(wch);
}
}
}
return wstr;
}
// Method Description:
// - get delimiter class for a position in the char row
// - used for double click selection and uia word navigation
// Arguments:
// - column: column to get text data for
// - wordDelimiters: the delimiters defined as a part of the DelimiterClass::DelimiterChar
// Return Value:
// - the delimiter class for the given char
const DelimiterClass CharRow::DelimiterClassAt(const size_t column, const std::wstring_view wordDelimiters) const
{
THROW_HR_IF(E_INVALIDARG, column >= _data.size());
const auto glyph = *GlyphAt(column).begin();
if (glyph <= UNICODE_SPACE)
{
return DelimiterClass::ControlChar;
}
else if (wordDelimiters.find(glyph) != std::wstring_view::npos)
{
return DelimiterClass::DelimiterChar;
}
else
{
return DelimiterClass::RegularChar;
}
}
UnicodeStorage& CharRow::GetUnicodeStorage() noexcept
{
return _pParent->GetUnicodeStorage();
}
const UnicodeStorage& CharRow::GetUnicodeStorage() const noexcept
{
return _pParent->GetUnicodeStorage();
}
// Routine Description:
// - calculates the key used by the given column of the char row to store glyph data in UnicodeStorage
// Arguments:
// - column - the column to generate the key for
// Return Value:
// - the COORD key for data access from UnicodeStorage for the column
COORD CharRow::GetStorageKey(const size_t column) const noexcept
{
return { gsl::narrow<SHORT>(column), _pParent->GetId() };
}
// Routine Description:
// - Updates the pointer to the parent row (which might change if we shuffle the rows around)
// Arguments:
// - pParent - Pointer to the parent row
void CharRow::UpdateParent(ROW* const pParent)
{
_pParent = FAIL_FAST_IF_NULL(pParent);
}