Implement allocation tracking

0x002A · Oct 28, 2021 · 3a25612 · 3a25612
1 parent f7748ec
commit 3a25612
Show file tree

Hide file tree

Showing 15 changed files with 435 additions and 56 deletions.
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -63,6 +63,11 @@ add_subdirectory("lib${LIBNAME}")
 file(GLOB_RECURSE SOURCES RELATIVE ${CMAKE_SOURCE_DIR} "src/*.cpp")
 add_executable(${TARGET} ${SOURCES})
 
+option(ENABLE_TRACK_MEMORY_USAGE "Enable tracking of memory allocations" OFF)
+if (ENABLE_TRACK_MEMORY_USAGE)
+    target_compile_definitions(${TARGET} PRIVATE TRACK_MEMORY_USAGE)
+endif ()
+
 option(ENABLE_BUILD_FAT_EXE "Enable linking dependencies statically" OFF)
 if (ENABLE_BUILD_FAT_EXE)
     target_link_options(${TARGET} PRIVATE -static)

diff --git a/include/ms/IO.h b/include/ms/IO.h
@@ -0,0 +1,50 @@
+// -*- C++ -*-
+//===---------------------------------------------------------------------------------------------------------------==//
+//
+// Copyright (C) 2021 Kevin Klein
+// This file is part of MuCHSALSA <https://github.com/0x002A/MuCHSALSA>.
+//
+// MuCHSALSA is free software: you can redistribute it and/or modify it under the terms of the GNU General
+// Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any
+// later version.
+//
+// MuCHSALSA is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the
+// implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more
+// details.
+//
+// You should have received a copy of the GNU General Public License along with MuCHSALSA.
+// If not, see <http://www.gnu.org/licenses/>.
+//
+// SPDX-License-Identifier: GPL-3.0-or-later
+//
+//===---------------------------------------------------------------------------------------------------------------==//
+
+#ifndef INCLUDED_MUCHSALSA_IO
+#define INCLUDED_MUCHSALSA_IO
+
+#pragma once
+
+#include <cstddef>
+#include <cstdint>
+#include <cstdio>
+
+namespace muchsalsa {
+
+// =====================================================================================================================
+//                                                    FREE FUNCTIONS
+// =====================================================================================================================
+
+/**
+ * Reads an entire line from a file, storing the address of the buffer containing the text using the supplied pointer.
+ * The buffer will be null-terminated and will contain the newline character, if one was found.
+ *
+ * @param pPtrBuffer a pointer to a pointer receiving the buffer address
+ * @param pSizeBuffer a pointer to a std::size_t representing the size of the supplied buffer
+ * @param pFile a pointer to a FILE handle required for accessing the file
+ * @return The number of characters read or -1 in the event of a failure
+ */
+int64_t readline(char **pPtrBuffer, std::size_t *pSizeBuffer, std::FILE *pFile);
+
+} // namespace muchsalsa
+
+#endif // INCLUDED_MUCHSALSA_IO
diff --git a/include/ms/Util.h b/include/ms/Util.h
@@ -30,7 +30,7 @@
 #include <type_traits>
 #include <utility>
 
-#define LB_UNUSED(x) (void)(x)
+#define MS_UNUSED(x) (void)(x)
 
 namespace muchsalsa::util {
 

diff --git a/include/ms/graph/Graph.h b/include/ms/graph/Graph.h
@@ -957,7 +957,7 @@ auto GraphUtil::getShortestPath(
     }
 
     for (auto const &[neighbor, pEdge] : _getReachableVertices(*pGraph, pVertex)) {
-      LB_UNUSED(pEdge);
+      MS_UNUSED(pEdge);
       auto const *const pNeighbor = pGraph->getVertex(neighbor);
 
       auto const distNeighbor = distances[pVertex] + 1;

diff --git a/libms/src/BlastFileAccessor.cpp b/libms/src/BlastFileAccessor.cpp
@@ -21,9 +21,10 @@
 
 #include "BlastFileAccessor.h"
 
-#include <cstdlib>
 #include <gsl/pointers>
 
+#include "IO.h"
+
 namespace muchsalsa {
 
 // =====================================================================================================================
@@ -51,12 +52,12 @@ std::string BlastFileAccessor::getLine(int64_t const &offset) {
 
   std::fseek(m_pBlastFile.get(), offset, SEEK_SET);
 
-  char *      pLine      = nullptr;
-  std::size_t bufferSize = 0;
-  auto        ret        = getline(&pLine, &bufferSize, m_pBlastFile.get());
+  char       *pLine      = nullptr;
+  std::size_t sizeBuffer = 0;
+  auto        ret        = readline(&pLine, &sizeBuffer, m_pBlastFile.get());
 
   auto result = [=]() {
-    if (ret != -1 && bufferSize > 0) {
+    if (ret != -1 && sizeBuffer > 0) {
       auto line = std::string(pLine);
       line.pop_back();
 
@@ -66,31 +67,27 @@ std::string BlastFileAccessor::getLine(int64_t const &offset) {
     return std::string();
   }();
 
-  if (pLine) {
-    std::free(pLine); // NOLINT
-  }
+  operator delete(pLine); // NOLINT
 
   return result;
 }
 
 // PRIVATE CLASS METHODS
 
 void BlastFileAccessor::_buildIndex() {
-  char *      pLine      = nullptr;
-  std::size_t bufferSize = 0;
+  char       *pLine      = nullptr;
+  std::size_t sizeBuffer = 0;
   auto        offset     = std::ftell(m_pBlastFile.get());
-  auto        ret        = getline(&pLine, &bufferSize, m_pBlastFile.get());
+  auto        ret        = readline(&pLine, &sizeBuffer, m_pBlastFile.get());
 
   while (ret != -1) {
     m_offsets.push_back(offset);
 
     offset = std::ftell(m_pBlastFile.get());
-    ret    = getline(&pLine, &bufferSize, m_pBlastFile.get());
+    ret    = readline(&pLine, &sizeBuffer, m_pBlastFile.get());
   }
 
-  if (pLine) {
-    std::free(pLine); // NOLINT
-  }
+  operator delete(pLine); // NOLINT
 }
 
 } // namespace muchsalsa

diff --git a/libms/src/IO.cpp b/libms/src/IO.cpp
@@ -0,0 +1,101 @@
+// -*- C++ -*-
+//===---------------------------------------------------------------------------------------------------------------==//
+//
+// Copyright (C) 2021 Kevin Klein
+// This file is part of MuCHSALSA <https://github.com/0x002A/MuCHSALSA>.
+//
+// MuCHSALSA is free software: you can redistribute it and/or modify it under the terms of the GNU General
+// Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any
+// later version.
+//
+// MuCHSALSA is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the
+// implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for more
+// details.
+//
+// You should have received a copy of the GNU General Public License along with MuCHSALSA.
+// If not, see <http://www.gnu.org/licenses/>.
+//
+// SPDX-License-Identifier: GPL-3.0-or-later
+//
+//===---------------------------------------------------------------------------------------------------------------==//
+
+#include "IO.h"
+
+#include <cstring>
+
+namespace muchsalsa {
+
+// =====================================================================================================================
+//                                                        HELPER
+// =====================================================================================================================
+
+namespace {
+
+void *realloc(void *pOldMem, std::size_t sizeOld, std::size_t sizeNew) {
+  auto *pNewMem = ::operator new(sizeNew);
+
+  if (!pOldMem || !pNewMem) {
+    return nullptr;
+  }
+
+  std::memcpy(pNewMem, pOldMem, sizeOld);
+
+  ::operator delete(pOldMem);
+
+  return pNewMem;
+}
+
+} // unnamed namespace
+
+// =====================================================================================================================
+//                                                    FREE FUNCTIONS
+// =====================================================================================================================
+
+int64_t readline(char **pPtrBuffer, std::size_t *pSizeBuffer, std::FILE *pFile) {
+  if (*pPtrBuffer == nullptr || *pSizeBuffer == 0) {
+    *pSizeBuffer = 120;
+
+    *pPtrBuffer = static_cast<char *>(::operator new(*pSizeBuffer));
+    if (*pPtrBuffer == nullptr) {
+      return -1;
+    }
+  }
+
+  auto *pCurrentBufferPos = *pPtrBuffer;
+  auto *pLastBufferPos    = *pPtrBuffer + *pSizeBuffer - 1; // NOLINT
+  auto  c                 = std::fgetc(pFile);
+  while (c != -1) {
+    *pCurrentBufferPos++ = static_cast<char>(c); // NOLINT
+    if (c == '\n') {
+      *pCurrentBufferPos = '\0';
+      return pCurrentBufferPos - *pPtrBuffer;
+    }
+    if (pCurrentBufferPos + 1 > pLastBufferPos) { // NOLINT
+      std::size_t sizeNewBuffer = *pSizeBuffer * 2;
+      auto        charsInBuffer = static_cast<std::size_t>(pCurrentBufferPos - *pPtrBuffer);
+
+      auto *const pNewBuffer = static_cast<char *>(realloc(*pPtrBuffer, charsInBuffer, sizeNewBuffer));
+      if (!pNewBuffer) {
+        return -1;
+      }
+
+      *pPtrBuffer       = pNewBuffer;
+      *pSizeBuffer      = sizeNewBuffer;
+      pLastBufferPos    = pNewBuffer + sizeNewBuffer; // NOLINT
+      pCurrentBufferPos = pNewBuffer + charsInBuffer; // NOLINT
+    }
+
+    c = std::fgetc(pFile);
+  }
+
+  if (std::feof(pFile) && pCurrentBufferPos != *pPtrBuffer) {
+    *pCurrentBufferPos = '\0';
+    return pCurrentBufferPos - *pPtrBuffer;
+  }
+
+  return -1;
+}
+
+} // namespace muchsalsa
+
+// ---------------------------------------------------- END-OF-FILE ----------------------------------------------------
diff --git a/libms/src/SequenceAccessor.cpp b/libms/src/SequenceAccessor.cpp
@@ -23,13 +23,12 @@
 
 #include <algorithm>
 #include <cctype>
-#include <cstdlib>
 #include <cstring>
 #include <span>
-#include <unordered_map>
 #include <utility>
 #include <vector>
 
+#include "IO.h"
 #include "Registry.h"
 #include "Util.h"
 #include "threading/Job.h"
@@ -145,28 +144,28 @@ void SequenceAccessor::_buildNanoporeIdx(gsl::not_null<threading::Job const *> p
   auto const identifierDescline  = m_nanoporeFileIsFastQ ? FASTQ_IDENTIFIER_DESCLINE : FASTA_IDENTIFIER_DESCLINE;
   auto const identifierSplitline = m_nanoporeFileIsFastQ ? FASTQ_IDENTIFIER_SPLITLINE : FASTA_IDENTIFIER_DESCLINE;
 
-  char *      pLine       = nullptr;
-  std::size_t bufferSize  = 0;
-  auto        ret         = getline(&pLine, &bufferSize, m_pNanoporeSequenceFile.get());
+  char       *pLine       = nullptr;
+  std::size_t sizeBuffer  = 0;
+  auto        ret         = readline(&pLine, &sizeBuffer, m_pNanoporeSequenceFile.get());
   auto        offsetStart = std::ftell(m_pNanoporeSequenceFile.get());
   while (ret != -1) {
     if (*pLine == identifierDescline) {
       break;
     }
 
-    ret         = getline(&pLine, &bufferSize, m_pNanoporeSequenceFile.get());
+    ret         = readline(&pLine, &sizeBuffer, m_pNanoporeSequenceFile.get());
     offsetStart = std::ftell(m_pNanoporeSequenceFile.get());
   }
 
   while (*pLine == identifierDescline) {
-    std::span const spanLine(pLine, bufferSize);
+    std::span const spanLine(pLine, sizeBuffer);
     auto            sequenceId = std::string(spanLine.subspan(1).data());
     cleanSequenceId(sequenceId);
 
     auto lengthCurrentSequence = 0L;
 
     while (true) {
-      ret            = getline(&pLine, &bufferSize, m_pNanoporeSequenceFile.get());
+      ret            = readline(&pLine, &sizeBuffer, m_pNanoporeSequenceFile.get());
       auto offsetEnd = std::ftell(m_pNanoporeSequenceFile.get());
 
       if (ret == -1 || *pLine == identifierSplitline) {
@@ -180,41 +179,39 @@ void SequenceAccessor::_buildNanoporeIdx(gsl::not_null<threading::Job const *> p
     }
 
     while (ret != -1 && *pLine != identifierDescline) {
-      ret         = getline(&pLine, &bufferSize, m_pNanoporeSequenceFile.get());
+      ret         = readline(&pLine, &sizeBuffer, m_pNanoporeSequenceFile.get());
       offsetStart = std::ftell(m_pNanoporeSequenceFile.get());
     }
   }
 
-  if (pLine) {
-    std::free(pLine); // NOLINT
-  }
+  operator delete(pLine); // NOLINT
 
   std::any_cast<threading::WaitGroup *>(pJob->getParam(0))->done();
 }
 
 void SequenceAccessor::_buildIlluminaIdx(gsl::not_null<threading::Job const *> pJob) {
-  char *      pLine       = nullptr;
-  std::size_t bufferSize  = 0;
-  auto        ret         = getline(&pLine, &bufferSize, m_pIlluminaSequenceFile.get());
+  char       *pLine       = nullptr;
+  std::size_t sizeBuffer  = 0;
+  auto        ret         = readline(&pLine, &sizeBuffer, m_pIlluminaSequenceFile.get());
   auto        offsetStart = std::ftell(m_pIlluminaSequenceFile.get());
   while (ret != -1) {
     if (*pLine == FASTA_IDENTIFIER_DESCLINE) {
       break;
     }
 
-    ret         = getline(&pLine, &bufferSize, m_pIlluminaSequenceFile.get());
+    ret         = readline(&pLine, &sizeBuffer, m_pIlluminaSequenceFile.get());
     offsetStart = std::ftell(m_pIlluminaSequenceFile.get());
   }
 
   while (*pLine == FASTA_IDENTIFIER_DESCLINE) {
-    std::span const spanLine(pLine, bufferSize);
+    std::span const spanLine(pLine, sizeBuffer);
     auto            sequenceId = std::string(spanLine.subspan(1).data());
     cleanSequenceId(sequenceId);
 
     auto lengthCurrentSequence = 0L;
 
     while (true) {
-      ret            = getline(&pLine, &bufferSize, m_pIlluminaSequenceFile.get());
+      ret            = readline(&pLine, &sizeBuffer, m_pIlluminaSequenceFile.get());
       auto offsetEnd = std::ftell(m_pIlluminaSequenceFile.get());
 
       if (ret == -1 || *pLine == FASTA_IDENTIFIER_DESCLINE) {
@@ -228,9 +225,7 @@ void SequenceAccessor::_buildIlluminaIdx(gsl::not_null<threading::Job const *> p
     }
   }
 
-  if (pLine) {
-    std::free(pLine); // NOLINT
-  }
+  operator delete(pLine); // NOLINT
 
   std::any_cast<threading::WaitGroup *>(pJob->getParam(0))->done();
 }

diff --git a/libms/src/graph/Graph.cpp b/libms/src/graph/Graph.cpp
@@ -185,7 +185,7 @@ void GraphBase::_deleteVertex(gsl::not_null<Vertex const *> const pVertex, bool
 
   if (!hasBidirectionalEdges) {
     for (auto &[targetId, connectedVertices] : m_adjacencyList) {
-      LB_UNUSED(targetId);
+      MS_UNUSED(targetId);
 
       auto const iterPredecessor = connectedVertices.find(pVertex->getId());
       if (iterPredecessor != std::end(connectedVertices)) {
@@ -376,7 +376,7 @@ std::vector<muchsalsa::graph::Vertex const *> DiGraph::sortTopologically() const
 
     auto const successors = getSuccessors(pVertex);
     for (auto const &[targetId, pEdge] : successors) {
-      LB_UNUSED(pEdge);
+      MS_UNUSED(pEdge);
 
       auto const *pSuccessor = getVertex(targetId);