Cancel wikidiff2 when it takes too long.

mediawiki-php-wikidiff2

@@ -1 +1 @@
-Subproject commit e487d37c58bd2147eef3bdd2e65544f1465f839a
+Subproject commit d8f34d2e7281bcbfa66d28939c3d03f25352ef0a

pywikidiff2/pywikidiff2.cpp

@@ -1,6 +1,9 @@
 #define PY_SSIZE_T_CLEAN
 #include <Python.h>
 #include <vector>
+#include <thread>
+#include <chrono>
+#include <atomic>
 #include "structmember.h"
 #undef HAVE_CONFIG_H // this disables the php allocator
 #include "../mediawiki-php-wikidiff2/src/lib/Wikidiff2.h"
@@ -18,6 +21,7 @@ using wikidiff2::InlineJSONFormatter;
 typedef struct {
   PyObject_HEAD
   Wikidiff2::Config config;
+  bool timed_out;
 } pywikidiff2Obj;
 
 // here's where we set default configuration values
@@ -37,6 +41,8 @@ pywikidiff2_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
     self->config.wordsCacheCapacity = 10000;
     self->config.diffCacheCapacity = 10000;
     self->config.statsCacheCapacity = 100000;
+    self->config.cancelFlag = nullptr;
+    self->timed_out = false;
     return (PyObject *) self;
 }
 
@@ -97,23 +103,49 @@ static PyObject *pywikidiff2_finalSplitThreshold(pywikidiff2Obj *self, PyObject
   return PyFloat_FromDouble(self->config.finalSplitThreshold);
 }
 
+static PyObject *pywikidiff2_timed_out(pywikidiff2Obj *self, PyObject *Py_UNUSED(ignored)){
+  return PyBool_FromLong(self->timed_out);
+}
+
 
 static Wikidiff2::String char_to_string(char* cstr){
   Wikidiff2::String str(cstr, strlen(cstr));
   return str;
 }
 
-static Wikidiff2::String wikidiff2_inline_json_diff(pywikidiff2Obj *self, char* text1, char* text2){
+static Wikidiff2::String wikidiff2_inline_json_diff(pywikidiff2Obj *self, char* text1, char* text2, long timeout_ms){
+  std::atomic<bool> cancelFlag(false);
+  self->config.cancelFlag = &cancelFlag;
+  self->timed_out = false;
+
+  std::thread timerThread;
+  if (timeout_ms > 0) {
+    timerThread = std::thread([&cancelFlag, timeout_ms]() {
+      std::this_thread::sleep_for(std::chrono::milliseconds(timeout_ms));
+      cancelFlag.store(true, std::memory_order_relaxed);
+    });
+  }
+
   Wikidiff2::String ret;
   Py_BEGIN_ALLOW_THREADS
   Wikidiff2::String str1(text1, strlen(text1));
   Wikidiff2::String str2(text2, strlen(text2));
-  Wikidiff2 wikidiff2( *(&self->config));
+  Wikidiff2 wikidiff2(self->config);
   InlineJSONFormatter formatter;
   wikidiff2.addFormatter(formatter);
   wikidiff2.execute(str1, str2);
   ret = formatter.getResult().str();
   Py_END_ALLOW_THREADS
+
+  // Check if we timed out before stopping the timer thread
+  self->timed_out = cancelFlag.load(std::memory_order_relaxed);
+
+  if (timerThread.joinable()) {
+    cancelFlag.store(true, std::memory_order_relaxed);  // Stop timer if still waiting
+    timerThread.join();
+  }
+
+  self->config.cancelFlag = nullptr;
   return ret;
 } 
 
@@ -186,7 +218,7 @@ static PyObject *pywikidiff2_inline_json_diff_sequence(pywikidiff2Obj *self, PyO
       text = new char[text_len];
       strcpy(text, input_texts_str[i]);
     }
-    Wikidiff2::String diff_str = wikidiff2_inline_json_diff(self, last_text, text);
+    Wikidiff2::String diff_str = wikidiff2_inline_json_diff(self, last_text, text, 0);
     PyObject* py_diff = PyUnicode_FromFormat("%s",diff_str.c_str());
     PyList_SetItem(result_list, i-1, py_diff);
   }
@@ -196,30 +228,33 @@ static PyObject *pywikidiff2_inline_json_diff_sequence(pywikidiff2Obj *self, PyO
 }
 
 
-static PyObject *pywikidiff2_inline_json_diff(pywikidiff2Obj *self, PyObject *args, PyObject *kwargs){ 
+static PyObject *pywikidiff2_inline_json_diff(pywikidiff2Obj *self, PyObject *args, PyObject *kwargs){
-  static char* kwdlist[] = {"text1", "text2", "numContextLines", NULL};
+  static char* kwdlist[] = {"text1", "text2", "numContextLines", "timeout_ms", NULL};
   char* text1;
   char* text2;
-  long numContextLines = NULL;
+  int numContextLines = 0;
-  if (!PyArg_ParseTupleAndKeywords(args, kwargs, "ss|$i",
+  long timeout_ms = 0;
+  if (!PyArg_ParseTupleAndKeywords(args, kwargs, "ss|il",
 				   kwdlist,
 				   &text1,
 				   &text2,
-				   &numContextLines
+				   &numContextLines,
+				   &timeout_ms
 				   )){
     PyErr_SetString(PyExc_ValueError, "Error in arguments to inline_json_diff");
     return NULL;
   }
-  if(numContextLines != NULL){
+  if(numContextLines != 0){
     self->config.numContextLines = numContextLines;
-  } 
+  }
-  Wikidiff2::String ret = wikidiff2_inline_json_diff(self, text1, text2);
+  Wikidiff2::String ret = wikidiff2_inline_json_diff(self, text1, text2, timeout_ms);
   return PyUnicode_FromFormat("%s",ret.c_str());
 };
 
 static PyMethodDef pywikidiff2_methods[] = {
   {"inline_json_diff", (PyCFunction) pywikidiff2_inline_json_diff, METH_VARARGS | METH_KEYWORDS, "run wikidiff 2 on text1 and text2"},
   {"inline_json_diff_sequence", (PyCFunction) pywikidiff2_inline_json_diff_sequence, METH_VARARGS | METH_KEYWORDS, "run wikidiff 2 on a series of texts"},
+  {"timed_out", (PyCFunction) pywikidiff2_timed_out, METH_NOARGS, "Whether the last diff timed out"},
   {"num_context_lines", (PyCFunction) pywikidiff2_numContextLines, METH_NOARGS, "number of equal lines to output in the context of a diff"},
   {"moved_line_threshold", (PyCFunction) pywikidiff2_movedLineThreshold, METH_NOARGS, "The minimum similarity a pair of lines must have to be detected as a moved line. If present, this overrides php.ini wikidiff2.moved_line_threshold"},
   {"change_threshold", (PyCFunction) pywikidiff2_changeThreshold, METH_NOARGS, "Changed lines with a similarity value below this threshold will be split into a deleted line and added line. This helps matching up moved lines in some cases."},