Skip to contentSkip to conversation

selftests/tpm2: Change exception handling to be Python 3 compatible

PATCH In review Jarkko Sakkinen 2 participants 9 messages Apr 12, 2020 – Apr 14, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 selftests/tpm2: Change exception handling to be Python 3 compatible

I need more time to fix all the byte array / string related stuff but it makes sense to fix the exceptions as it is fairly mechanical procedure:

1,$s/except \(.*\), \(.*\):/except \1(\2):/g

I.e. fix the low hanging fruit first and the rest later.

---
 tools/testing/selftests/tpm2/tpm2_tests.py | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
tools/testing/selftests/tpm2/tpm2_tests.py +6 -6
index 728be7c69b76..c3c06899e042 100644
@@ -65,7 +65,7 @@ class SmokeTest(unittest.TestCase):
         blob = self.client.seal(self.root_key, data, auth, None)
         try:
             result = self.client.unseal(self.root_key, blob, auth[:-1] + 'B', None)
-        except ProtocolError, e:
+        except ProtocolError(e):
             rc = e.rc
 
         self.assertEqual(rc, tpm2.TPM2_RC_AUTH_FAIL)
@@ -119,7 +119,7 @@ class SmokeTest(unittest.TestCase):
             self.client.policy_password(handle)
 
             result = self.client.unseal(self.root_key, blob, auth, handle)
-        except ProtocolError, e:
+        except ProtocolError(e):
             rc = e.rc
             self.client.flush_context(handle)
         except:
@@ -136,7 +136,7 @@ class SmokeTest(unittest.TestCase):
         rc = 0
         try:
             blob = self.client.seal(self.root_key, data, auth, None)
-        except ProtocolError, e:
+        except ProtocolError(e):
             rc = e.rc
 
         self.assertEqual(rc, tpm2.TPM2_RC_SIZE)
@@ -152,7 +152,7 @@ class SmokeTest(unittest.TestCase):
                               0xDEADBEEF)
 
             self.client.send_cmd(cmd)
-        except IOError, e:
+        except IOError(e):
             rejected = True
         except:
             pass
@@ -212,7 +212,7 @@ class SmokeTest(unittest.TestCase):
             self.client.tpm.write(cmd)
             rsp = self.client.tpm.read()
 
-        except IOError, e:
+        except IOError(e):
             # read the response
             rsp = self.client.tpm.read()
             rejected = True
@@ -283,7 +283,7 @@ class SpaceTest(unittest.TestCase):
         rc = 0
         try:
             space1.send_cmd(cmd)
-        except ProtocolError, e:
+        except ProtocolError(e):
             rc = e.rc
 
         self.assertEqual(rc, tpm2.TPM2_RC_COMMAND_CODE |
Signature · 1 line
2.25.1
#1 lore raw

Hi Jarkkon,

Quoted 2 lines · Jarkko Sakkinen

+ except ProtocolError(e):

Should this not be

except ProtocolError as e:

?

Cheers,
Ezra.

#2 lore raw
Quoted 8 lines · Ezra Buehler

Hi Jarkkon,

Quoted 2 lines · Jarkko Sakkinen

+ except ProtocolError(e):

Should this not be

except ProtocolError as e:

Unless there is a functional difference, does it matter?

/Jarkko

#3 lore raw

Hi Jarkko,

Quoted 7 lines · Jarkko Sakkinen
Quoted 5 lines · Ezra Buehler

Hi Jarkkon,

Quoted 2 lines · Jarkko Sakkinen

+ except ProtocolError(e):

Quoted 7 lines
Quoted 3 lines

Should this not be

except ProtocolError as e:

Unless there is a functional difference, does it matter?

/Jarkko

Well, your patch confuses me a lot. It looks to me like you are passing the undefined e variable to the constructor.

When I run flake8 on it I get following error (among others):

F821 undefined name 'e'

What I suggested is the standard syntax:
https://docs.python.org/3/tutorial/errors.html

Did you test this? You should get an error as soon as an exception occurs.

Cheers,
Ezra.

#4 lore raw
Quoted 16 lines · Ezra Buehler

Hi Jarkko,

Quoted 13 lines · Jarkko Sakkinen
Quoted 8 lines · Ezra Buehler

Hi Jarkkon,

Quoted 2 lines · Jarkko Sakkinen

+ except ProtocolError(e):

Should this not be

except ProtocolError as e:

Unless there is a functional difference, does it matter?

Quoted 8 lines
Quoted 1 line

/Jarkko

Well, your patch confuses me a lot. It looks to me like you are passing the undefined e variable to the constructor.

When I run flake8 on it I get following error (among others):

F821 undefined name 'e'

I don't know what flake8 is.

Quoted 2 lines

What I suggested is the standard syntax:
https://docs.python.org/3/tutorial/errors.html

It passed the Python 3 interpreter.

Quoted 2 lines

Did you test this? You should get an error as soon as an exception occurs.

Yes. Interpreter did not complain. I did not know that the language is broken that way that you have to exercise the code path to get a syntax error.

/Jarkko

#5 lore raw

Hi Jarkko,

Quoted 20 lines · Jarkko Sakkinen
Quoted 18 lines · Ezra Buehler

Hi Jarkko,

Quoted 14 lines · Jarkko Sakkinen
Quoted 8 lines · Ezra Buehler

Hi Jarkkon,

Quoted 2 lines · Jarkko Sakkinen

+ except ProtocolError(e):

Should this not be

except ProtocolError as e:

Unless there is a functional difference, does it matter?

/Jarkko

Quoted 8 lines
Quoted 6 lines

Well, your patch confuses me a lot. It looks to me like you are passing the undefined e variable to the constructor.

When I run flake8 on it I get following error (among others):

F821 undefined name 'e'

I don't know what flake8 is.

https://flake8.pycqa.org/en/latest/

Quoted 4 lines
Quoted 2 lines

What I suggested is the standard syntax:
https://docs.python.org/3/tutorial/errors.html

It passed the Python 3 interpreter.

That is because it is technically valid syntax.

Quoted 6 lines
Quoted 2 lines

Did you test this? You should get an error as soon as an exception occurs.

Yes. Interpreter did not complain. I did not know that the language is broken that way that you have to exercise the code path to get a syntax error.

That is due to the dynamic nature of Python. You won’t get a syntax error. You will get an exception:

NameError: name 'e' is not defined

Python has to assume that e might be defined at runtime. However, style checkers will complain.

Quoted 2 lines

/Jarkko

Cheers,
Ezra.

#6 lore raw
Quoted 48 lines · Ezra Buehler

Hi Jarkko,

Quoted 20 lines · Jarkko Sakkinen
Quoted 18 lines · Ezra Buehler

Hi Jarkko,

Quoted 14 lines · Jarkko Sakkinen
Quoted 8 lines · Ezra Buehler

Hi Jarkkon,

Quoted 2 lines · Jarkko Sakkinen

+ except ProtocolError(e):

Should this not be

except ProtocolError as e:

Unless there is a functional difference, does it matter?

/Jarkko

Quoted 8 lines
Quoted 6 lines

Well, your patch confuses me a lot. It looks to me like you are passing the undefined e variable to the constructor.

When I run flake8 on it I get following error (among others):

F821 undefined name 'e'

I don't know what flake8 is.

https://flake8.pycqa.org/en/latest/

Quoted 4 lines
Quoted 2 lines

What I suggested is the standard syntax:
https://docs.python.org/3/tutorial/errors.html

It passed the Python 3 interpreter.

That is because it is technically valid syntax.

Quoted 6 lines
Quoted 2 lines

Did you test this? You should get an error as soon as an exception occurs.

Yes. Interpreter did not complain. I did not know that the language is broken that way that you have to exercise the code path to get a syntax error.

Quoted 7 lines

That is due to the dynamic nature of Python. You won’t get a syntax error. You will get an exception:

NameError: name 'e' is not defined

Python has to assume that e might be defined at runtime. However, style checkers will complain.

OK, I'm aware about the dynamic nature but in this case it is somewhat counter intuitive since it is part of the exception clause. You'd except the Python interpreter to complain.

So, is Flake8 like the standard to be used?

/Jarkko

#7 lore raw
Quoted 3 lines · Jarkko Sakkinen

OK, I'm aware about the dynamic nature but in this case it is somewhat counter intuitive since it is part of the exception clause. You'd except the Python interpreter to complain.

I agree.

Quoted 1 line

So, is Flake8 like the standard to be used?

Pretty much, yes. There is also Pylint though. Among other things, they both check for PEP 8 (official) coding style compliance.

Cheers,
Ezra.

Quoted 66 lines · Jarkko Sakkinen
Quoted 48 lines · Ezra Buehler

Hi Jarkko,

Quoted 20 lines · Jarkko Sakkinen
Quoted 18 lines · Ezra Buehler

Hi Jarkko,

Quoted 14 lines · Jarkko Sakkinen
Quoted 8 lines · Ezra Buehler

Hi Jarkkon,

Quoted 2 lines · Jarkko Sakkinen

+ except ProtocolError(e):

Should this not be

except ProtocolError as e:

Unless there is a functional difference, does it matter?

/Jarkko

Quoted 8 lines
Quoted 6 lines

Well, your patch confuses me a lot. It looks to me like you are passing the undefined e variable to the constructor.

When I run flake8 on it I get following error (among others):

F821 undefined name 'e'

I don't know what flake8 is.

https://flake8.pycqa.org/en/latest/

Quoted 4 lines
Quoted 2 lines

What I suggested is the standard syntax:
https://docs.python.org/3/tutorial/errors.html

It passed the Python 3 interpreter.

That is because it is technically valid syntax.

Quoted 6 lines
Quoted 2 lines

Did you test this? You should get an error as soon as an exception occurs.

Yes. Interpreter did not complain. I did not know that the language is broken that way that you have to exercise the code path to get a syntax error.

Quoted 7 lines

That is due to the dynamic nature of Python. You won’t get a syntax error. You will get an exception:

NameError: name 'e' is not defined

Python has to assume that e might be defined at runtime. However, style checkers will complain.

OK, I'm aware about the dynamic nature but in this case it is somewhat counter intuitive since it is part of the exception clause. You'd except the Python interpreter to complain.

So, is Flake8 like the standard to be used?

/Jarkko

#8 lore raw
Quoted 11 lines · Ezra Buehler
Quoted 3 lines · Jarkko Sakkinen

OK, I'm aware about the dynamic nature but in this case it is somewhat counter intuitive since it is part of the exception clause. You'd except the Python interpreter to complain.

I agree.

Quoted 1 line

So, is Flake8 like the standard to be used?

Pretty much, yes. There is also Pylint though. Among other things, they both check for PEP 8 (official) coding style compliance.

Thank you for taking time explaining all this. I'll make sure to include these to my process when I update my test from now on.

/Jarkko