Add a test for RefScanSink and clean up the code

Issue #5322.
This commit is contained in:
Eelco Dolstra 2021-10-04 14:29:42 +02:00
parent ef34fd0656
commit 77ebbc9f54
5 changed files with 105 additions and 30 deletions

View file

@ -4,6 +4,7 @@ makefiles = \
src/libutil/local.mk \
src/libutil/tests/local.mk \
src/libstore/local.mk \
src/libstore/tests/local.mk \
src/libfetchers/local.mk \
src/libmain/local.mk \
src/libexpr/local.mk \

View file

@ -11,11 +11,13 @@
namespace nix {
static unsigned int refLength = 32; /* characters */
static size_t refLength = 32; /* characters */
static void search(const unsigned char * s, size_t len,
StringSet & hashes, StringSet & seen)
static void search(
std::string_view s,
StringSet & hashes,
StringSet & seen)
{
static std::once_flag initialised;
static bool isBase32[256];
@ -25,7 +27,7 @@ static void search(const unsigned char * s, size_t len,
isBase32[(unsigned char) base32Chars[i]] = true;
});
for (size_t i = 0; i + refLength <= len; ) {
for (size_t i = 0; i + refLength <= s.size(); ) {
int j;
bool match = true;
for (j = refLength - 1; j >= 0; --j)
@ -35,7 +37,7 @@ static void search(const unsigned char * s, size_t len,
break;
}
if (!match) continue;
string ref((const char *) s + i, refLength);
std::string ref(s.substr(i, refLength));
if (hashes.erase(ref)) {
debug(format("found reference to '%1%' at offset '%2%'")
% ref % i);
@ -46,30 +48,23 @@ static void search(const unsigned char * s, size_t len,
}
struct RefScanSink : Sink
void RefScanSink::operator () (std::string_view data)
{
StringSet hashes;
StringSet seen;
/* 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;
s.append(data.data(), refLength);
search(s, hashes, seen);
string tail;
search(data, hashes, seen);
RefScanSink() { }
void operator () (std::string_view data) override
{
/* 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. */
string s = tail + std::string(data, 0, refLength);
search((const unsigned char *) s.data(), s.size(), hashes, seen);
search((const unsigned char *) data.data(), data.size(), hashes, seen);
size_t tailLen = data.size() <= refLength ? data.size() : refLength;
tail = std::string(tail, tail.size() < refLength - tailLen ? 0 : tail.size() - (refLength - tailLen));
tail.append({data.data() + data.size() - tailLen, tailLen});
}
};
auto tailLen = std::min(data.size(), refLength);
auto rest = refLength - tailLen;
if (rest < tail.size())
tail = tail.substr(tail.size() - rest);
tail.append(data.data() + data.size() - tailLen, tailLen);
}
std::pair<StorePathSet, HashResult> scanForReferences(
@ -87,23 +82,24 @@ StorePathSet scanForReferences(
const Path & path,
const StorePathSet & refs)
{
RefScanSink refsSink;
TeeSink sink { refsSink, toTee };
StringSet hashes;
std::map<std::string, StorePath> backMap;
for (auto & i : refs) {
std::string hashPart(i.hashPart());
auto inserted = backMap.emplace(hashPart, i).second;
assert(inserted);
refsSink.hashes.insert(hashPart);
hashes.insert(hashPart);
}
/* Look for the hashes in the NAR dump of the path. */
RefScanSink refsSink(std::move(hashes));
TeeSink sink { refsSink, toTee };
dumpPath(path, sink);
/* Map the hashes found back to their store paths. */
StorePathSet found;
for (auto & i : refsSink.seen) {
for (auto & i : refsSink.getResult()) {
auto j = backMap.find(i);
assert(j != backMap.end());
found.insert(j->second);

View file

@ -9,6 +9,24 @@ std::pair<StorePathSet, HashResult> scanForReferences(const Path & path, const S
StorePathSet scanForReferences(Sink & toTee, const Path & path, const StorePathSet & refs);
class RefScanSink : public Sink
{
StringSet hashes;
StringSet seen;
std::string tail;
public:
RefScanSink(StringSet && hashes) : hashes(hashes)
{ }
StringSet & getResult()
{ return seen; }
void operator () (std::string_view data) override;
};
struct RewritingSink : Sink
{
std::string from, to, prev;

View file

@ -0,0 +1,15 @@
check: libstore-tests_RUN
programs += libstore-tests
libstore-tests_DIR := $(d)
libstore-tests_INSTALL_DIR :=
libstore-tests_SOURCES := $(wildcard $(d)/*.cc)
libstore-tests_CXXFLAGS += -I src/libstore -I src/libutil
libstore-tests_LIBS = libstore
libstore-tests_LDFLAGS := $(GTEST_LIBS)

View file

@ -0,0 +1,45 @@
#include "references.hh"
#include <gtest/gtest.h>
namespace nix {
TEST(references, scan)
{
std::string hash1 = "dc04vv14dak1c1r48qa0m23vr9jy8sm0";
std::string hash2 = "zc842j0rz61mjsp3h3wp5ly71ak6qgdn";
{
RefScanSink scanner(StringSet{hash1});
auto s = "foobar";
scanner(s);
ASSERT_EQ(scanner.getResult(), StringSet{});
}
{
RefScanSink scanner(StringSet{hash1});
auto s = "foobar" + hash1 + "xyzzy";
scanner(s);
ASSERT_EQ(scanner.getResult(), StringSet{hash1});
}
{
RefScanSink scanner(StringSet{hash1, hash2});
auto s = "foobar" + hash1 + "xyzzy" + hash2;
scanner(((std::string_view) s).substr(0, 10));
scanner(((std::string_view) s).substr(10, 5));
scanner(((std::string_view) s).substr(15, 5));
scanner(((std::string_view) s).substr(20));
ASSERT_EQ(scanner.getResult(), StringSet({hash1, hash2}));
}
{
RefScanSink scanner(StringSet{hash1, hash2});
auto s = "foobar" + hash1 + "xyzzy" + hash2;
for (auto & i : s)
scanner(std::string(1, i));
ASSERT_EQ(scanner.getResult(), StringSet({hash1, hash2}));
}
}
}