mirror of
https://github.com/NixOS/nix.git
synced 2025-11-16 23:42:43 +01:00
f3e1c47f47
The short answer for why we need to do this is so we can consistently do `#include "nix/..."`. Without this change, there are ways to still make that work, but they are hacky, and they have downsides such as making it harder to make sure headers from the wrong Nix library (e..g. `libnixexpr` headers in `libnixutil`) aren't being used. The C API alraedy used `nix_api_*`, so its headers are *not* put in subdirectories accordingly. Progress on #7876 We resisted doing this for a while because it would be annoying to not have the header source file pairs close by / easy to change file path/name from one to the other. But I am ameliorating that with symlinks in the next commit.
139 lines
3.5 KiB
C++
139 lines
3.5 KiB
C++
#include "nix/references.hh"
|
|
#include "nix/hash.hh"
|
|
#include "nix/archive.hh"
|
|
|
|
#include <map>
|
|
#include <cstdlib>
|
|
#include <mutex>
|
|
#include <algorithm>
|
|
|
|
|
|
namespace nix {
|
|
|
|
|
|
static size_t refLength = 32; /* characters */
|
|
|
|
|
|
static void search(
|
|
std::string_view s,
|
|
StringSet & hashes,
|
|
StringSet & seen)
|
|
{
|
|
static std::once_flag initialised;
|
|
static bool isBase32[256];
|
|
std::call_once(initialised, [](){
|
|
for (unsigned int i = 0; i < 256; ++i) isBase32[i] = false;
|
|
for (unsigned int i = 0; i < nix32Chars.size(); ++i)
|
|
isBase32[(unsigned char) nix32Chars[i]] = true;
|
|
});
|
|
|
|
for (size_t i = 0; i + refLength <= s.size(); ) {
|
|
int j;
|
|
bool match = true;
|
|
for (j = refLength - 1; j >= 0; --j)
|
|
if (!isBase32[(unsigned char) s[i + j]]) {
|
|
i += j + 1;
|
|
match = false;
|
|
break;
|
|
}
|
|
if (!match) continue;
|
|
std::string ref(s.substr(i, refLength));
|
|
if (hashes.erase(ref)) {
|
|
debug("found reference to '%1%' at offset '%2%'", ref, i);
|
|
seen.insert(ref);
|
|
}
|
|
++i;
|
|
}
|
|
}
|
|
|
|
|
|
void RefScanSink::operator () (std::string_view data)
|
|
{
|
|
/* It's possible that a reference spans the previous and current
|
|
fragment, so search in the concatenation of the tail of the
|
|
previous fragment and the start of the current fragment. */
|
|
auto s = tail;
|
|
auto tailLen = std::min(data.size(), refLength);
|
|
s.append(data.data(), tailLen);
|
|
search(s, hashes, seen);
|
|
|
|
search(data, hashes, seen);
|
|
|
|
auto rest = refLength - tailLen;
|
|
if (rest < tail.size())
|
|
tail = tail.substr(tail.size() - rest);
|
|
tail.append(data.data() + data.size() - tailLen, tailLen);
|
|
}
|
|
|
|
|
|
RewritingSink::RewritingSink(const std::string & from, const std::string & to, Sink & nextSink)
|
|
: RewritingSink({{from, to}}, nextSink)
|
|
{
|
|
}
|
|
|
|
RewritingSink::RewritingSink(const StringMap & rewrites, Sink & nextSink)
|
|
: rewrites(rewrites), nextSink(nextSink)
|
|
{
|
|
std::string::size_type maxRewriteSize = 0;
|
|
for (auto & [from, to] : rewrites) {
|
|
assert(from.size() == to.size());
|
|
maxRewriteSize = std::max(maxRewriteSize, from.size());
|
|
}
|
|
this->maxRewriteSize = maxRewriteSize;
|
|
}
|
|
|
|
void RewritingSink::operator () (std::string_view data)
|
|
{
|
|
std::string s(prev);
|
|
s.append(data);
|
|
|
|
s = rewriteStrings(s, rewrites);
|
|
|
|
prev = s.size() < maxRewriteSize
|
|
? s
|
|
: maxRewriteSize == 0
|
|
? ""
|
|
: std::string(s, s.size() - maxRewriteSize + 1, maxRewriteSize - 1);
|
|
|
|
auto consumed = s.size() - prev.size();
|
|
|
|
pos += consumed;
|
|
|
|
if (consumed) nextSink(s.substr(0, consumed));
|
|
}
|
|
|
|
void RewritingSink::flush()
|
|
{
|
|
if (prev.empty()) return;
|
|
pos += prev.size();
|
|
nextSink(prev);
|
|
prev.clear();
|
|
}
|
|
|
|
HashModuloSink::HashModuloSink(HashAlgorithm ha, const std::string & modulus)
|
|
: hashSink(ha)
|
|
, rewritingSink(modulus, std::string(modulus.size(), 0), hashSink)
|
|
{
|
|
}
|
|
|
|
void HashModuloSink::operator () (std::string_view data)
|
|
{
|
|
rewritingSink(data);
|
|
}
|
|
|
|
HashResult HashModuloSink::finish()
|
|
{
|
|
rewritingSink.flush();
|
|
|
|
/* Hash the positions of the self-references. This ensures that a
|
|
NAR with self-references and a NAR with some of the
|
|
self-references already zeroed out do not produce a hash
|
|
collision. FIXME: proof. */
|
|
for (auto & pos : rewritingSink.matches)
|
|
hashSink(fmt("|%d", pos));
|
|
|
|
auto h = hashSink.finish();
|
|
return {h.first, rewritingSink.pos};
|
|
}
|
|
|
|
}
|