bpo-30721: Show correct syntax hint in Py3 when using Py2 redirection syntax #2345

CuriousLearner · 2017-06-23T05:27:14Z

This is a work in progress for issue: http://bugs.python.org/issue30721

@ncoghlan I'm not sure how should I get the value that is being entered on the command line. What I know is it would be in PyObject *w, but I'm unable to get the right method from the doc to read the buffer.

So, for now, I've hard-coded the value just to make the test case pass. Could you please help me with the former problem and give me a pointer on it.

Also, after writing this patch, I realize, that binary_op is a generic function, and this code should be placed in function binop_type_error. Would you please guide me about these things?

Thanks!

https://bugs.python.org/issue30721

mention-bot · 2017-06-23T05:27:18Z

@CuriousLearner, thanks for your PR! By analyzing the history of the files in this pull request, we identified @serhiy-storchaka, @benjaminp and @nascheme to be potential reviewers.

ncoghlan

Comments inline regarding some lower level details. I haven't looked at whether or not this is the right place to make the change yet, though.

ncoghlan · 2017-06-23T06:11:34Z

Objects/abstract.c

@@ -819,6 +819,21 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
    PyObject *result = binary_op1(v, w, op_slot);
    if (result == Py_NotImplemented) {
        Py_DECREF(result);
+
+        if (strcmp(op_name, ">>") == 0 && \


I need to re-read the whole file myself to see if there may be a better place to inject this check, but in the mean time, at least compare against the numeric op_slot here, rather than the string value.

ncoghlan · 2017-06-23T06:24:00Z

Objects/abstract.c

+            PyErr_Format(PyExc_TypeError,
+                 "unsupported operand type(s) for %.100s: "
+                 "'%.100s' and '%.100s'. Did you mean \"print(%s, "
+                 "file={:100!r}).format(%s))\"",


Don't include an actual format call in the error message - that was just what I used in the issue tracker to represent PyErr_Format pllaceholders. Instead, you'll want to use the %R format code to invoke PyObject_Repr on the operands directly:

PyErr_Format(PyExc_TypeError, "unsupported operand type(s) for %.100s: " "'%.100s' and '%.100s'. Did you mean \"print(%.100R, " "file={.100R})\"", op_name, v->ob_type->tp_name, w->ob_type->tp_name, w, v);

In the general case, you won't be able to recreate the exact text of the original input, as that's no longer available by the time we get here - we only have a reference to the final runtime objects.

We might decide it's worthwhile adding a special-case check for the RHS being exactly PySys_GetObject("stderr") and displaying file=sys.stderr in that case.

Alternatively, we could dispense with the idea of interpolating values entirely, and just use the hardcoded output "Did you mean \"print(<message>, file=<output_stream>)\""

I'm leaning towards the "hardcoded placeholders" option, as it occurs to me that we don't actually have access to the message to be displayed when raising the exception:

print >> sys.stderr, 1, 2, 3

The LHS passed to >> is print, and the RHS is sys.stderr. I think that's why I just had <expr> as a placeholder in my example on the issue tracker.

However, the repr of a standard stream isn't particular helpful in this context either:

>>> sys.stderr <_io.TextIOWrapper name='<stderr>' mode='w' encoding='UTF-8'>

The generic <output_stream> placeholder would be clearer than displaying that.

So yeah, forget about interpolating the inputs at all, and just append a string with generic placeholders in it.

CuriousLearner · 2017-06-23T08:36:19Z

@ncoghlan I just noticed, I cannot move this snippet to binop_type_error function if I need to check for op_slot because that is not accessible there.

So, I formatted the error message in the binary_op function and let it return NULL as binop_type_error function does.

I've made the changes to include generic statement, but I think if there is a way to access op_slot with op_name, then this patch could be simplified. What are your thoughts on this?

terryjreedy · 2017-07-24T22:57:59Z

test_string_with_stream_redirection (test.test_print.TestPy2MigrationHint) ... FAIL

FAIL: test_string_with_stream_redirection (test.test_print.TestPy2MigrationHint)

Traceback (most recent call last):
File "C:\projects\cpython\lib\test\test_print.py", line 165, in test_string_with_stream_redirection
'file=<output_stream>)', str(context.exception))
AssertionError: 'Did you mean "print(, file=<output_stream>)' not found in "unsupported operand type(s) for >>: 'builtin_function_or_method' and '_io.StringIO'"

CuriousLearner · 2017-07-25T10:23:16Z

@terryjreedy The test passes on my local setup. I'm not sure why this is happening on CI.

I will have a look at it.

CuriousLearner · 2017-07-25T10:25:38Z

@terryjreedy Here is the result of test run on Mac OS

Run tests sequentially
0:00:00 load avg: 2.19 [1/1] test_print
1 test OK.

Total duration: 58 ms
Tests result: SUCCESS

terryjreedy · 2017-07-25T18:06:14Z

@CuriousLearner Travis, where it also passed, run linux. Appveyor, where it failed runs Windows. I asked on the issue for a Windows expert to take a look.

ncoghlan · 2017-07-26T14:52:46Z

Objects/abstract.c

@@ -819,6 +819,20 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
    PyObject *result = binary_op1(v, w, op_slot);
    if (result == Py_NotImplemented) {
        Py_DECREF(result);
+
+        if (op_slot == 96 && \


This should be op_slot == NB_SLOT(nb_rshift)

CuriousLearner · 2017-07-31T07:08:21Z

@ncoghlan @terryjreedy Please have a look now. The issue is resolved now.

serhiy-storchaka · 2017-07-27T04:52:37Z

Objects/abstract.c

@@ -819,6 +819,20 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
    PyObject *result = binary_op1(v, w, op_slot);
    if (result == Py_NotImplemented) {
        Py_DECREF(result);
+
+        if (op_slot == 96 && \
+                strcmp(v->ob_type->tp_name, "builtin_function_or_method") == 0) {


Don't test type's name. Test the print object itself.

Aye, you'll want to do something like:

_Py_IDENTIFIER(print); builtins = PyEval_GetBuiltins(); print_ref = _PyDict_GetItemId(builtins, &PyId_print);

and then check for v == print_ref.

(Only don't lay it out exactly like that, adjust it to fit the code - the above is just to show the key APIs you need)

ncoghlan · 2017-07-31T10:31:19Z

Objects/abstract.c

@@ -819,6 +819,20 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
    PyObject *result = binary_op1(v, w, op_slot);
    if (result == Py_NotImplemented) {
        Py_DECREF(result);
+
+        if (op_slot == 96 && \
+                strcmp(v->ob_type->tp_name, "builtin_function_or_method") == 0) {


Aye, you'll want to do something like:

_Py_IDENTIFIER(print); builtins = PyEval_GetBuiltins(); print_ref = _PyDict_GetItemId(builtins, &PyId_print);

and then check for v == print_ref.

(Only don't lay it out exactly like that, adjust it to fit the code - the above is just to show the key APIs you need)

ncoghlan · 2017-07-31T10:33:11Z

Objects/abstract.c

@@ -819,6 +819,20 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
    PyObject *result = binary_op1(v, w, op_slot);
    if (result == Py_NotImplemented) {
        Py_DECREF(result);
+
+        if (op_slot == NB_SLOT(nb_rshift) && \
+                strcmp(v->ob_type->tp_name, "builtin_function_or_method") == 0) {


Moving the review comment about checking for exactly the print builtin to the latest commit:

_Py_IDENTIFIER(print); builtins = PyEval_GetBuiltins(); print_ref = _PyDict_GetItemId(builtins, &PyId_print);

and then check for v == print_ref.

(Only don't lay it out exactly like that, adjust it to fit the code - the above is just to show the key APIs you need)

Yes, but since this code would be slow and complex, I think it can be replaced by something like

PyCFunction_Check(v) && PyCFunction_GET_FUNCTION(v) == builtin_print

or

PyCFunction_Check(v) && strcmp(((PyCFunctionObject *)v) -> m_ml -> ml_name, "print") == 0

Fair point. Since builtin_print is a static symbol in bltinmodule, your second suggestion looks good to me.

There are many ways to get around this constraint. Export builtin_print or implement the check in bltinmodule.c and export the checking code for using in abstract.c.

But testing the function name may be enough.

Right, that was my reasoning too - while we could make the actual symbol available here, "It's a function called 'print' implemented in C" is a robust enough check, and it means the C level changes can stay contained to this module.

serhiy-storchaka · 2017-07-31T11:02:56Z

Lib/test/test_print.py

+
+        self.assertIn('Did you mean "print(<message>, '
+                'file=<output_stream>)', str(context.exception))
+


Add following tests:

max >> sys.stderr.

print >> with right argument implemented __rrshift__.

Hi @serhiy-storchaka

Can you please explain a little about the above cases?

For 1, As far as I understand the max function when redirected to sys.stderr should not cause the Did you mean ... statement to be executed.

So, a test case can be

def test_stream_redirection_using_max(self): import sys python2_print_str = 'max >> sys.stderr' with self.assertRaises(TypeError) as context: exec(python2_print_str) self.assertNotIn('Did you mean "print(<message>, ' 'file=<output_stream>)', str(context.exception))

For the 2, I don't quite understand if we just want to test print >> with the current patch. In that case again the Did you mean... statement would not be executed.

Is this what we want to test here?

Please correct me.

Thanks!

@CuriousLearner Correct for the first case. For the second case, you'll want to implement a dummy class that does something like this:

class OverrideRRShift: def __rrshift__(self, lhs): return 42 # Force result independent of LHS self.assertEqual(print >> OverrideRRShift(), 42)

That will specifically ensure that the special casing doesn't trigger when the rshift operation produces a sensible answer.

It would also be good to test that print >> 5 does produce the new exception message (that covers the case where the RHS implements __rrshift__ but returns NotImplemented)

@ncoghlan Can you please take a look at the changes?

ncoghlan

This is starting to look pretty good to me, just one suggestion for another test case, as well as some suggestions for rearranging and renaming the existing test cases a bit.

ncoghlan · 2017-08-18T05:15:24Z

Lib/test/test_print.py

@@ -155,7 +155,7 @@ def test_string_with_excessive_whitespace(self):

        self.assertIn('print("Hello World", end=" ")', str(context.exception))

-    def test_string_with_stream_redirection(self):
+    def test_print_string_with_stream_redirection(self):


Suggested test name: test_stream_redirection_hint_for_py2_migration

ncoghlan · 2017-08-18T05:16:32Z

Lib/test/test_print.py

@@ -164,6 +164,32 @@ def test_string_with_stream_redirection(self):
        self.assertIn('Did you mean "print(<message>, '
                'file=<output_stream>)', str(context.exception))

+    def test_print_with_stream_redirection_using_max(self):


Suggested test name: test_stream_redirection_hint_is_specific_to_print

Prompted by that, I'd also suggest another test case: test_stream_redirection_hint_is_specific_to_rshift that checks:

print << sys.stderr

doesn't produce the new hint in its error message.

ncoghlan · 2017-08-18T05:19:12Z

Lib/test/test_print.py

+
+        self.assertEqual(print >> OverrideRRShift(), 0)
+
+        python2_print_str = 'print >> 42'


This is a situation where we expect to see the hint, so it would fit better up in the first test case.

Add a comment about what case is covered by this test. It looks as a duplicate of the first case.

ncoghlan · 2017-08-18T05:21:47Z

Lib/test/test_print.py

+        with self.assertRaises(TypeError) as context:
+            exec(python2_print_str)
+
+        self.assertNotIn('Did you mean "print(<message>, '


I'd shorten the test string here to just 'Did you mean'

ncoghlan · 2017-08-18T05:22:56Z

Lib/test/test_print.py

@@ -155,7 +155,7 @@ def test_string_with_excessive_whitespace(self):

        self.assertIn('print("Hello World", end=" ")', str(context.exception))

-    def test_string_with_stream_redirection(self):
+    def test_print_string_with_stream_redirection(self):
        import sys
        python2_print_str = 'print >> sys.stderr, "message"'


This is legal Python 3 syntax, so there's no need to use exec for these test cases - you can just put the expression in the body of the with statement.

@ncoghlan But that is just a string, we would need to execute it in the context manager, no?

Yes, that's what I'm saying - the code that checked for the custom SyntaxError needed to be tested with exec, because putting it outside a string would cause the entire module to fail to compile. That isn't the case here, so you can just do print >> sys.stderr, "message" directly in the body of the context manager (and similarly for the other cases).

ncoghlan · 2017-08-18T05:24:02Z

Lib/test/test_print.py

+        self.assertNotIn('Did you mean "print(<message>, '
+                'file=<output_stream>)', str(context.exception))
+
+    def test_print_with_stream_redirection_using_rrshift(self):


Suggested test name: test_no_stream_redirection_hint_when_rshift_succeeds

ncoghlan · 2017-08-18T05:28:46Z

Objects/abstract.c

@@ -821,7 +821,7 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
        Py_DECREF(result);

        if (op_slot == NB_SLOT(nb_rshift) && \
-                strcmp(v->ob_type->tp_name, "builtin_function_or_method") == 0) {
+                PyCFunction_Check(v) && strcmp(((PyCFunctionObject *)v) -> m_ml -> ml_name, "print") == 0) {


When working with user supplied strings, we generally prefer strncmp to strcmp, which would give:

strncmp(((PyCFunctionObject *)v) -> m_ml -> ml_name, "print", 6)

for the comparison subexpression.

What are reasons for using strncmp() instead of strcmp() here? "print" is NUL terminated, the comparison is limited to 6 chars in any case.

bedevere-bot · 2017-08-18T05:30:49Z

A Python core developer, ncoghlan, has requested some changes be
made to your pull request before we can consider merging it. If you
could please address their requests along with any other requests in
other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment
on this pull request containing the phrase I didn't expect the Spanish Inquisition!
I will then notify ncoghlan along with any other core developers
who have left a review that you're ready for them to take another look
at this pull request.

serhiy-storchaka · 2017-08-18T05:41:17Z

Lib/test/test_print.py

+        self.assertIn('Did you mean "print(<message>, '
+                'file=<output_stream>)', str(context.exception))
+
+    def test_print_with_stream_redirection_using_max(self):


I think it would be better to use the single test method for all these tests. Use empty lines and/or comments for separating them.

That also works - the suggested test_stream_redirection_hint_for_py2_migration name is sufficiently broad to cover all of them, and then the comment on each case can explain which particular aspect it is checking.

Sure, I'll do that.

serhiy-storchaka · 2017-08-18T05:42:10Z

Lib/test/test_print.py

+
+    def test_print_with_stream_redirection_using_rrshift(self):
+        import sys
+


Too much empty lines. Use empty lines only for separating different tests in one method.

serhiy-storchaka · 2017-08-18T05:43:15Z

Lib/test/test_print.py

+        class OverrideRRShift:
+
+            def __rrshift__(self, lhs):
+                return 0 # Force result independent of LHS


Use something more unique. Like 42 or 'spam'.

serhiy-storchaka · 2017-08-18T05:46:53Z

Objects/abstract.c

@@ -819,6 +819,20 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
    PyObject *result = binary_op1(v, w, op_slot);
    if (result == Py_NotImplemented) {
        Py_DECREF(result);
+
+        if (op_slot == NB_SLOT(nb_rshift) && \
+                PyCFunction_Check(v) && strcmp(((PyCFunctionObject *)v) -> m_ml -> ml_name, "print") == 0) {


Move strcmp to the next line. Align PyCFunction_Check and strcmp to op_slot.

According to new rules of PEP 7 the opening { after multiline condition should be at new line.

serhiy-storchaka · 2017-08-18T05:47:37Z

Objects/abstract.c

@@ -819,6 +819,20 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
    PyObject *result = binary_op1(v, w, op_slot);
    if (result == Py_NotImplemented) {
        Py_DECREF(result);
+
+        if (op_slot == NB_SLOT(nb_rshift) && \


\ is not needed.

serhiy-storchaka · 2017-08-18T05:50:15Z

Lib/test/test_print.py

+
+        self.assertEqual(print >> OverrideRRShift(), 0)
+
+        python2_print_str = 'print >> 42'


Add a comment about what case is covered by this test. It looks as a duplicate of the first case.

ncoghlan

Almost there :)

The NEWS entry needs to be adjusted to follow the new process for that, and I spotted a not-quite-right comment in the tests that I missed last time.

ncoghlan · 2017-08-18T08:27:01Z

Lib/test/test_print.py

-        self.assertNotIn('Did you mean "print(<message>, '
-                'file=<output_stream>)', str(context.exception))
+            print << sys.stderr
+        self.assertNotIn('Did you mean', str(context.exception))

        # Test stream redirection with right argument implemented __rrshift__


Sorry, I missed this comment on my last review: it should say something like "Ensure right operand implementing __rrshift__ still works"

ncoghlan · 2017-08-18T08:29:23Z

Misc/NEWS

@@ -10,6 +10,9 @@ What's New in Python 3.7.0 alpha 1?
 Core and Builtins
 -----------------

+- bpo-30721: ``print`` now shows correct usage hint for using Python 2
+  redirection syntax. Patch by Sanyam Khurana.
+


We don't edit Misc/NEWS directly any more. Instead, you'll want to add a blurb entry under https://github.com/python/cpython/tree/master/Misc/NEWS.d/next/Core%20and%20Builtins

See the existing entries and https://docs.python.org/devguide/committing.html#what-s-new-and-news-entries for more details

Blurb tool looks great! I faced a lot of merge conflicts earlier while rebasing my patches from the latest upstream changes.

This would greatly solve the problem.

bedevere-bot · 2017-08-18T08:31:48Z

A Python core developer, ncoghlan, has requested some changes be
made to your pull request before we can consider merging it. If you
could please address their requests along with any other requests in
other reviews from core developers that would be appreciated.

Once you have made the requested changes, please leave a comment
on this pull request containing the phrase I didn't expect the Spanish Inquisition!
I will then notify ncoghlan along with any other core developers
who have left a review that you're ready for them to take another look
at this pull request.

serhiy-storchaka · 2017-08-18T09:08:35Z

Lib/test/test_print.py

@@ -155,6 +156,39 @@ def test_string_with_excessive_whitespace(self):

        self.assertIn('print("Hello World", end=" ")', str(context.exception))

+    def test_stream_redirection_hint_for_py2_migration(self):
+


This one empty line is redundant.

serhiy-storchaka · 2017-08-18T09:18:06Z

Objects/abstract.c

@@ -821,7 +821,7 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)
        Py_DECREF(result);

        if (op_slot == NB_SLOT(nb_rshift) && \
-                strcmp(v->ob_type->tp_name, "builtin_function_or_method") == 0) {
+                PyCFunction_Check(v) && strcmp(((PyCFunctionObject *)v) -> m_ml -> ml_name, "print") == 0) {


What are reasons for using strncmp() instead of strcmp() here? "print" is NUL terminated, the comparison is limited to 6 chars in any case.

CuriousLearner · 2017-08-18T09:38:17Z

@serhiy-storchaka using strncmp was suggested by Nick in this comment

ncoghlan · 2017-08-18T09:49:43Z

Right, that was my suggestion, but I was getting my string comparison algorithms mixed up - strcmp doesn't include CPython's length-based pre-check (since without an O(1) length query it would just waste time), so it's OK to just let it terminate when it hits the end of the "print" literal. strncmp is only necessary when both inputs are potentially arbitrary length.

So yeah, you should change it back to what you had originally - sorry about that :)

ncoghlan · 2017-08-18T09:51:27Z

Aside from needing to revert my less-than-helpful strncmp suggestion though, this looks good to me now :)

CuriousLearner · 2017-08-18T09:57:01Z

Alright, fixed everything! :)

I didn't expect the Spanish Inquisition!

bedevere-bot · 2017-08-18T09:57:04Z

Nobody expects the Spanish Inquisition!

@serhiy-storchaka, @ncoghlan: please review the changes made to this pull request.

ncoghlan · 2017-08-18T09:59:12Z

Objects/abstract.c

@@ -822,7 +822,7 @@ binary_op(PyObject *v, PyObject *w, const int op_slot, const char *op_name)

        if (op_slot == NB_SLOT(nb_rshift) &&
            PyCFunction_Check(v) &&
-            strncmp(((PyCFunctionObject *)v)->m_ml->ml_name, "print", 6) == 0)
+            strcmp(((PyCFunctionObject *)v) -> m_ml -> ml_name, "print") == 0)


No spaces needed around ->.

CuriousLearner · 2017-08-18T10:05:41Z

I didn't expect the Spanish Inquisition!

bedevere-bot · 2017-08-18T10:05:43Z

Nobody expects the Spanish Inquisition!

@serhiy-storchaka, @ncoghlan: please review the changes made to this pull request.

ncoghlan

This looks good to me now, so over to @serhiy-storchaka for the final review :)

CuriousLearner · 2017-08-18T10:47:54Z

Thanks @serhiy-storchaka @ncoghlan :) 🥇

ncoghlan · 2017-08-18T10:53:26Z

Thanks @CuriousLearner for persevering with the change! I foresee a lot of folks making the Py2->Py3 migration benefiting from your work over the next few years :)

… syntax (python#2345)

the-knights-who-say-ni added the CLA signed label Jun 23, 2017

ncoghlan requested changes Jun 23, 2017

View reviewed changes

CuriousLearner added 2 commits July 12, 2017 23:18

bpo-30721: WIP for showing hint in Py3 when using Py2 redirection syntax

b26d729

bpo-30721: Address review comments

aa5cc11

the-knights-who-say-ni added CLA not signed and removed CLA signed labels Jul 12, 2017

CuriousLearner force-pushed the fix-issue30721 branch from 66d0529 to aa5cc11 Compare July 12, 2017 18:14

terryjreedy removed the CLA not signed label Jul 24, 2017

the-knights-who-say-ni added the CLA signed label Jul 24, 2017

python deleted a comment from the-knights-who-say-ni Jul 24, 2017

ncoghlan reviewed Jul 26, 2017

View reviewed changes

bpo-30721: Address review comments

7db3e71

CuriousLearner changed the title ~~bpo-30721: [WIP] for showing correct syntax hint in Py3 when using Py2 redirection syntax~~ bpo-30721: Show correct syntax hint in Py3 when using Py2 redirection syntax Jul 31, 2017

serhiy-storchaka requested changes Jul 31, 2017

View reviewed changes

ncoghlan requested changes Jul 31, 2017

View reviewed changes

ncoghlan reviewed Jul 31, 2017

View reviewed changes

serhiy-storchaka requested changes Jul 31, 2017

View reviewed changes

bpo-30721: Address review comments

6eaf005

ncoghlan requested changes Aug 18, 2017

View reviewed changes

bedevere-bot added the awaiting changes label Aug 18, 2017

serhiy-storchaka reviewed Aug 18, 2017

View reviewed changes

bpo-30721: Address review comments

7665115

ncoghlan requested changes Aug 18, 2017

View reviewed changes

bedevere-bot added awaiting changes and removed awaiting change review labels Aug 18, 2017

serhiy-storchaka reviewed Aug 18, 2017

View reviewed changes

bpo-30721: Adds NEWS entry and minor fixes

c43e23c

CuriousLearner added 2 commits August 18, 2017 15:23

bpo-30721: Fixes rST news entry

bad1d7b

bpo-30721: Reverts strncmp to strcmp

74eff16

bedevere-bot added awaiting change review and removed awaiting changes labels Aug 18, 2017

ncoghlan reviewed Aug 18, 2017

View reviewed changes

CuriousLearner added 2 commits August 18, 2017 15:29

bpo-30721: Removes manual entry to MISC/NEWS

d1a9641

bpo-30721: Removes spaces around pointers

185050e

ncoghlan approved these changes Aug 18, 2017

View reviewed changes

bedevere-bot added awaiting merge and removed awaiting change review labels Aug 18, 2017

serhiy-storchaka approved these changes Aug 18, 2017

View reviewed changes

serhiy-storchaka merged commit 5e2eb35 into python:master Aug 18, 2017

CuriousLearner mentioned this pull request Aug 18, 2017

bpo-30721: Addresses minor fixes of adding ? to redirection syntax #3131

Merged

ncoghlan pushed a commit to ncoghlan/cpython that referenced this pull request Aug 19, 2017

bpo-30721: Show correct syntax hint in Py3 when using Py2 redirection…

b1e3512

… syntax (python#2345)

Mariatta removed the awaiting merge label Aug 29, 2017


		self.assertIn('Did you mean "print(<message>, '
		'file=<output_stream>)', str(context.exception))


		self.assertEqual(print >> OverrideRRShift(), 0)

		python2_print_str = 'print >> 42'


		def test_print_with_stream_redirection_using_rrshift(self):
		import sys

		@@ -155,6 +156,39 @@ def test_string_with_excessive_whitespace(self):

		self.assertIn('print("Hello World", end=" ")', str(context.exception))

		def test_stream_redirection_hint_for_py2_migration(self):

bpo-30721: Show correct syntax hint in Py3 when using Py2 redirection syntax #2345

bpo-30721: Show correct syntax hint in Py3 when using Py2 redirection syntax #2345

Conversation

CuriousLearner commented Jun 23, 2017 • edited by bedevere-bot Loading

mention-bot commented Jun 23, 2017

ncoghlan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

CuriousLearner commented Jun 23, 2017

terryjreedy commented Jul 24, 2017

test_string_with_stream_redirection (test.test_print.TestPy2MigrationHint) ... FAIL

FAIL: test_string_with_stream_redirection (test.test_print.TestPy2MigrationHint)

CuriousLearner commented Jul 25, 2017

CuriousLearner commented Jul 25, 2017 • edited Loading

terryjreedy commented Jul 25, 2017

Choose a reason for hiding this comment

CuriousLearner commented Jul 31, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

serhiy-storchaka Jul 31, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ncoghlan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

serhiy-storchaka Aug 18, 2017 • edited Loading

Choose a reason for hiding this comment

bedevere-bot commented Aug 18, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ncoghlan left a comment

Choose a reason for hiding this comment

ncoghlan Aug 18, 2017 • edited by serhiy-storchaka Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

bedevere-bot commented Aug 18, 2017

Choose a reason for hiding this comment

serhiy-storchaka Aug 18, 2017 • edited Loading

Choose a reason for hiding this comment

CuriousLearner commented Aug 18, 2017

ncoghlan commented Aug 18, 2017

ncoghlan commented Aug 18, 2017

CuriousLearner commented Aug 18, 2017

bedevere-bot commented Aug 18, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

CuriousLearner commented Aug 18, 2017

bedevere-bot commented Aug 18, 2017

ncoghlan left a comment

Choose a reason for hiding this comment

CuriousLearner commented Aug 18, 2017

ncoghlan commented Aug 18, 2017

CuriousLearner commented Jun 23, 2017 •

edited by bedevere-bot

Loading

CuriousLearner commented Jul 25, 2017 •

edited

Loading

serhiy-storchaka Jul 31, 2017 •

edited

Loading

serhiy-storchaka Aug 18, 2017 •

edited

Loading

ncoghlan Aug 18, 2017 •

edited by serhiy-storchaka

Loading

serhiy-storchaka Aug 18, 2017 •

edited

Loading