From 13b0ebee7cdb1e4d200b3c40d0ec9440f198a1d4 Mon Sep 17 00:00:00 2001
Message-Id: <13b0ebee7cdb1e4d200b3c40d0ec9440f198a1d4.1554886141.git.pmatilai@redhat.com>
From: Panu Matilainen <pmatilai@redhat.com>
Date: Wed, 10 Apr 2019 11:24:44 +0300
Subject: [PATCH] Monkey-patch .decode() method to our strings as a temporary
 compat crutch

As a temporary crutch to support faster deployment of the sane
string behavior on python3, monkey-patch a decode method into all
strings we return. This seems to be enough to fix practically all
API users who have already adapted to the long-standing broken API
on Python 3. API users compatible with both Python 2 and 3 never needed
this anyway. Issue a warning with pointer to the relevant bug when the
fake decode() method is used to alert users to the issue.

This is certainly an evil thing to do and will be removed as soon as
the critical users have been fixed to work with the new, corrected
behavior.
---
 python/rpm/__init__.py |  3 +++
 python/rpmmodule.c     |  1 +
 python/rpmsystem-py.h  | 22 ++++++++++++++++++++--
 3 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/python/rpm/__init__.py b/python/rpm/__init__.py
index 54728bbd4..6d69eda7b 100644
--- a/python/rpm/__init__.py
+++ b/python/rpm/__init__.py
@@ -61,6 +61,9 @@ except ImportError:
 # backwards compatibility + give the same class both ways
 ts = TransactionSet
 
+def _fakedecode(self, encoding='utf-8', errors='strict'):
+    warnings.warn("decode() called on unicode string, see https://bugzilla.redhat.com/show_bug.cgi?id=1693751", UnicodeWarning, stacklevel=2)
+    return self
 
 def headerLoad(*args, **kwds):
     """DEPRECATED! Use rpm.hdr() instead."""
diff --git a/python/rpmmodule.c b/python/rpmmodule.c
index 05032edc7..2a76cfbd0 100644
--- a/python/rpmmodule.c
+++ b/python/rpmmodule.c
@@ -28,6 +28,7 @@
  */
 
 PyObject * pyrpmError;
+PyObject * fakedecode = NULL;
 
 static PyObject * archScore(PyObject * self, PyObject * arg)
 {
diff --git a/python/rpmsystem-py.h b/python/rpmsystem-py.h
index 25938464a..803da0fc1 100644
--- a/python/rpmsystem-py.h
+++ b/python/rpmsystem-py.h
@@ -19,12 +19,30 @@
 #define PyInt_AsSsize_t PyLong_AsSsize_t
 #endif
 
+PyObject * fakedecode;
+
 static inline PyObject * utf8FromString(const char *s)
 {
 /* In Python 3, we return all strings as surrogate-escaped utf-8 */
 #if PY_MAJOR_VERSION >= 3
-    if (s != NULL)
-	return PyUnicode_DecodeUTF8(s, strlen(s), "surrogateescape");
+    if (s != NULL) {
+	PyObject *o = PyUnicode_DecodeUTF8(s, strlen(s), "surrogateescape");
+	/* fish the fake decode function from python side if not done yet */
+	if (fakedecode == NULL) {
+	    PyObject *n = PyUnicode_FromString("rpm");
+	    PyObject *m = PyImport_Import(n);
+	    PyObject *md = PyModule_GetDict(m);
+	    fakedecode = PyDict_GetItemString(md, "_fakedecode");
+	    Py_DECREF(m);
+	    Py_DECREF(n);
+	}
+	if (fakedecode && o) {
+	    Py_INCREF(fakedecode);
+	    /* monkey-patch it into the string object as "decode" */
+	    PyDict_SetItemString(Py_TYPE(o)->tp_dict, "decode", fakedecode);
+	}
+	return o;
+    }
 #else
     if (s != NULL)
 	return PyBytes_FromString(s);
-- 
2.20.1