malloc/memset even less

more buffers that can be uninitialized and on the stack. small
difference, but still worth doing.

before:

Benchmark 1: nix eval --raw --impure --expr 'with import <nixpkgs/nixos> {}; system'
  Time (mean ± σ):      6.963 s ±  0.011 s    [User: 5.330 s, System: 1.421 s]
  Range (min … max):    6.943 s …  6.974 s    10 runs

after:

Benchmark 1: nix eval --raw --impure --expr 'with import <nixpkgs/nixos> {}; system'
  Time (mean ± σ):      6.952 s ±  0.015 s    [User: 5.294 s, System: 1.452 s]
  Range (min … max):    6.926 s …  6.974 s    10 runs

src/libutil/archive.cc

@@ -140,7 +140,7 @@ static void parseContents(ParseSink & sink, Source & source, const Path & path)
     sink.preallocateContents(size);
 
     uint64_t left = size;
-    std::vector<char> buf(65536);
+    std::array<char, 65536> buf;
 
     while (left) {
         checkInterrupt();

src/libutil/file-system.cc

@@ -307,7 +307,7 @@ void writeFile(const Path & path, Source & source, mode_t mode, bool sync)
     if (!fd)
         throw SysError("opening file '%1%'", path);
 
-    std::vector<char> buf(64 * 1024);
+    std::array<char, 64 * 1024> buf;
 
     try {
         while (true) {

src/libutil/serialise.cc

@@ -82,7 +82,7 @@ void Source::operator () (std::string_view data)
 void Source::drainInto(Sink & sink)
 {
     std::string s;
-    std::vector<char> buf(8192);
+    std::array<char, 8192> buf;
     while (true) {
         size_t n;
         try {