Fix possible segfault when importing multiple schemas

It turns out that parseDiskFile expects the imports array to last the
lifetime of the SchemaParser object.

Also, optimized for the case where the imports haven't changed, so we're
not allocating a crazy amount of these imports arrays.

capnp/lib/capnp.pyx

@@ -2979,6 +2979,23 @@ class _EnumModule(object):
         for name, val in schema.enumerants.items():
             setattr(self, name, val)
 
+cdef class _StringArrayPtr:
+    cdef StringPtr * thisptr
+    cdef object parent
+    cdef size_t size
+
+    def __cinit__(self, size_t size, parent):
+        self.size = size
+        self.thisptr = <StringPtr *>malloc(sizeof(StringPtr) * size)
+        self.parent = parent
+
+    def __dealloc__(self):
+        free(self.thisptr)
+
+    cdef ArrayPtr[StringPtr] asArrayPtr(self) except +reraise_kj_exception:
+        return ArrayPtr[StringPtr](self.thisptr, self.size)
+
+
 cdef class SchemaParser:
     """A class for loading Cap'n Proto schema files.
 
@@ -2986,26 +3003,34 @@ cdef class SchemaParser:
     """
     cdef C_SchemaParser * thisptr
     cdef public dict modules_by_id
+    cdef list _all_imports
+    cdef _StringArrayPtr _last_import_array
 
     def __cinit__(self):
         self.thisptr = new C_SchemaParser()
         self.modules_by_id = {}
+        self._all_imports = []
 
     def __dealloc__(self):
         del self.thisptr
 
     cpdef _parse_disk_file(self, displayName, diskPath, imports) except +reraise_kj_exception:
-        cdef StringPtr * importArray = <StringPtr *>malloc(sizeof(StringPtr) * len(imports))
+        cdef _StringArrayPtr importArray
 
-        for i in range(len(imports)):
-            importArray[i] = StringPtr(imports[i])
+        if self._last_import_array and self._last_import_array.parent == imports:
+            importArray = self._last_import_array
+        else:
+            importArray = _StringArrayPtr(len(imports), imports)
 
-        cdef ArrayPtr[StringPtr] importsPtr = ArrayPtr[StringPtr](importArray, <size_t>len(imports))
+            for i in range(len(imports)):
+                curr_import = imports[i]
+                importArray.thisptr[i] = StringPtr(curr_import, <size_t>len(curr_import))
+
+            self._all_imports.append(importArray)
+            self._last_import_array = importArray
 
         ret = _ParsedSchema()
-        ret._init_child(self.thisptr.parseDiskFile(displayName, diskPath, importsPtr))
-
-        free(importArray)
+        ret._init_child(self.thisptr.parseDiskFile(displayName, diskPath, importArray.asArrayPtr()))
 
         return ret