Dino Viehland wrote:

This is probably a dup of 12907 http://ironpython.codeplex.com/WorkItem/View.aspx?WorkItemId=12907

That being said the more test cases for this the better – this option isn’t really documented. Also if anyone can provide insight into exactly how this is supposed to behave that’d be great. But my guess is that we’ll have a tail of bugs around this until we can nail down the behavior and get it right. Anyway, I’ll take a look at fixing the existing issues here. Wanting to implement a REPL from Python seems like it’s the thing everyone likes to try J


I wonder if it isn't in fact caused by the fact that the repr of an IronPython syntax error doesn't change if the you add new lines to the source code you are compiling. As you can see the code below is actually comparing the error repr with newline added to see what *kind* of SyntaxError it is (horrible I know but this is what the code module does).

There is a codeplex issue for this specifically but I can't find it.

Michael

*From:* users-boun...@lists.ironpython.com [mailto:users-boun...@lists.ironpython.com] *On Behalf Of *Mike Krell
*Sent:* Monday, May 25, 2009 10:00 AM
*To:* Discussion of IronPython
*Subject:* [IronPython] IPython is breathing but there's a compile() problem

Now that 2.6B1 has frames support, I've started playing with IronPython under IPython again. I've managed to get a command prompt up (some modules are missing, but the only crucial one is codeop, which I stole from the standard distribution).

However, there's a problem with entering multiline code snippets interactively. With CPython, this looks like:

In [21]: if 1:
....: if 1:
....:

(The indentation looks wrong without a fixed-width font, but you get the idea.)

With IronPython, the second "if 1:" line blows up with a syntax error. This boils down to a difference in the way the compile() builtin works as used by the codeop module.

I've written this up as a bug at codeplex. Please vote for the bug here:

http://ironpython.codeplex.com/WorkItem/View.aspx?WorkItemId=22692

It would be awesome if we could have a good IronPython + IPython story before 2.6 is released!

Below are more details about the problem as described in the bug description.

Mike

--------------------------------------------------------------------------------

bug description at codeplex follows

compile() behaves differently than in CPython in the presence of incomplete multiline code snippets. Fixing this incompatiblity is necessary for running IronPython under IPython.

Here is a sample program illustrating the problem. The program is a modification of the code used in the standard codeop module by IPython to determine when to provide a continuation prompt for a multiline snippet.

--------------------------------------------------------------------------------

def testcompile(source, flags):

err = err1 = err2 = None
code = code1 = code2 = None

try:
code = compile(source, "dummy", "single", flags, 1)
except SyntaxError, err:
pass

try:
code1 = compile(source + "\n", "dummy", "single", flags, 1)
except SyntaxError, err1:
pass

try:
code2 = compile(source + "\n\n", "dummy", "single", flags, 1)
except SyntaxError, err2:
pass

print "for source = '%s' and flags = %d" % (source, flags),

if code:
print "Syntax valid"
elif not code1 and repr(err1) == repr(err2):
print "Syntax error!"
print
print "err1:", repr(err1)
print "err2:", repr(err2)
else:
print "Continue on next line"
print
print "err1:", repr(err1)
print "err2:", repr(err2)

print

# 0x200 is PyCF_DONT_IMPLY_DEDENT

testcompile("if 1:", 0x200)
testcompile("if 1:", 0)

testcompile("if 1:\n if 1:", 0x200)
testcompile("if 1:\n if 1:", 0)

--------------------------------------------------------------------------------

Under CPython (2.6.1) the output is:

--------------------------------------------------------------------------------

for source = 'if 1:' and flags = 512 Continue on next line

err1: SyntaxError('unexpected EOF while parsing', ('dummy', 1, 6, 'if 1:\n')) err2: IndentationError('expected an indented block', ('dummy', 2, 1, '\n'))

for source = 'if 1:' and flags = 0 Continue on next line

err1: SyntaxError('unexpected EOF while parsing', ('dummy', 1, 6, 'if 1:\n')) err2: IndentationError('expected an indented block', ('dummy', 2, 1, '\n'))

for source = 'if 1:
if 1:' and flags = 512 Continue on next line

err1: IndentationError('expected an indented block', ('dummy', 2, 8, ' if 1:\n')) err2: IndentationError('expected an indented block', ('dummy', 3, 1, '\n'))

for source = 'if 1:
if 1:' and flags = 0 Continue on next line

err1: IndentationError('expected an indented block', ('dummy', 2, 8, ' if 1:\n')) err2: IndentationError('expected an indented block', ('dummy', 3, 1, '\n'))

--------------------------------------------------------------------------------


In all cases the code correctly outputs "Continue on next line" since both snippets are incomplete but otherwise valid python.

For IronPython 2.6 Beta 1 the output is:

--------------------------------------------------------------------------------

for source = 'if 1:' and flags = 512 Continue on next line

err1: IndentationError("unexpected token '<eof>'", ('dummy', 2, 1, ''))
err2: IndentationError("unexpected token '<eof>'", ('dummy', 3, 1, ''))

for source = 'if 1:' and flags = 0 Continue on next line

err1: IndentationError("unexpected token '<eof>'", ('dummy', 2, 1, ''))
err2: IndentationError("unexpected token '<eof>'", ('dummy', 3, 1, ''))

for source = 'if 1:
if 1:' and flags = 512 Syntax error!

err1: IndentationError("unexpected token '<eof>'", ('dummy', 2, 8, ' if 1:\n')) err2: IndentationError("unexpected token '<eof>'", ('dummy', 2, 8, ' if 1:\n'))

for source = 'if 1:
if 1:' and flags = 0 Syntax error!

err1: IndentationError("unexpected token '<eof>'", ('dummy', 2, 8, ' if 1:\n')) err2: IndentationError("unexpected token '<eof>'", ('dummy', 2, 8, ' if 1:\n'))

--------------------------------------------------------------------------------

The second snippet is misinterpreted as being a syntax error instead of merely incomplete.

This is very similar to the issue described here:

http://www.voidspace.org.uk/python/weblog/arch_d7_2007_09_22.shtml#e834

But it apparently doesn't have to do with PyCF_DONT_IMPLY_DEDENT per se because the CPython output is the same whether this flag is passed or not.



------------------------------------------------------------------------

_______________________________________________
Users mailing list
Users@lists.ironpython.com
http://lists.ironpython.com/listinfo.cgi/users-ironpython.com


--
http://www.ironpythoninaction.com/
http://www.voidspace.org.uk/blog


_______________________________________________
Users mailing list
Users@lists.ironpython.com
http://lists.ironpython.com/listinfo.cgi/users-ironpython.com

Reply via email to