##// END OF EJS Templates
revlog: add a small cache of unfiltered chunk...
revlog: add a small cache of unfiltered chunk This can provides a massive boost to the reading of multiple revision and the computation of a valid delta chain. This greatly help operation like `hg log --patch`, delta computation (helping pull/unbundle), linkrev adjustment (helping copy tracing). A first round of benchmark for `hg log --patch --limit 1000` shows improvement in the 10-20% range on "small" repository like pypy or mercurial and large improvements (about 33%) for more complex ones like netbeans and mozilla's. These speeds up are consistent with the improvement to `hg pull` (from a server sending poor deltas) I saw benchmarking this last year. Further benchmark will be run during the freeze. I added some configuration in the experimental space to be able to further test the effect of various tuning for now. This feature should fit well in the "usage/resource profile" configuration that we should land next cycle. When it does not provides a benefit the overhead of the cache seem to be around 2%, a small price for the big improvement. In addition I believe we could shave most of this overhead with a more efficent lru implementation.

File last commit:

r44311:8766728d default
r52001:0250e450 default
Show More
fncache.cc
69 lines | 1.8 KiB | text/x-c | CppLexer
#include <Python.h>
#include <assert.h>
#include <stdlib.h>
#include <unistd.h>
#include "pyutil.h"
#include <iostream>
#include <string>
extern "C" {
static PYCODETYPE *code;
extern "C" int LLVMFuzzerInitialize(int *argc, char ***argv)
{
contrib::initpy(*argv[0]);
code = (PYCODETYPE *)Py_CompileString(R"py(
try:
for fn in (
parsers.isasciistr,
parsers.asciilower,
parsers.asciiupper,
parsers.encodedir,
parsers.pathencode,
parsers.lowerencode,
):
try:
fn(data)
except UnicodeDecodeError:
pass # some functions emit this exception
except AttributeError:
# pathencode needs hashlib, which fails to import because the time
# module fails to import. We should try and fix that some day, but
# for now we at least get coverage on non-hashencoded codepaths.
if fn != pathencode:
raise
# uncomment this for debugging exceptions
# except Exception as e:
# raise Exception('%r: %r' % (fn, e))
except Exception as e:
pass
# uncomment this print if you're editing this Python code
# to debug failures.
# print(e)
)py",
"fuzzer", Py_file_input);
if (!code) {
std::cerr << "failed to compile Python code!" << std::endl;
}
return 0;
}
int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size)
{
PyObject *mtext =
PyBytes_FromStringAndSize((const char *)Data, (Py_ssize_t)Size);
PyObject *locals = PyDict_New();
PyDict_SetItemString(locals, "data", mtext);
PyObject *res = PyEval_EvalCode(code, contrib::pyglobals(), locals);
if (!res) {
PyErr_Print();
}
Py_XDECREF(res);
Py_DECREF(locals);
Py_DECREF(mtext);
return 0; // Non-zero return values are reserved for future use.
}
}