Conversion error in converting Redhawk device from 1.8.4 to 1.9 - redhawksdr

Question

I am using Redhawk 1.9. I have created a Redhawk Device from the 1.8.4 IDE.
1.8.4 Redhawk Device (created with all default settings)
C++ implementation
Import existing project into 1.9 IDE
I read the Redhawk 1.9 Release notes on how to convert between 1.8.4 and 1.9 projects
*http://redhawksdr.github.io/Documentation/Release_Notes_1.9.0.pdf*
I import the default 1.8.4 device into the 1.9 IDE. I am able to run and build the 1.8.4 device in the 1.9 IDE. When I try to regenerate the code for 1.8.4 device, the IDE ask me if I want to upgrade to 1.9. The popup says "ConversionTestDevice uses deprecated code generators. Would you like to upgrade this project?". I decided to do an upgrade. I then get the following error message:
/usr/local/redhawk/core/bin/update_project returned with error code 1
Traceback (most recent call last):
File "/usr/local/redhawk/core/bin/update_project", line 222, in ?
if check_bulkio_input(compCpp):
File "/usr/local/redhawk/core/bin/update_project", line 105, in check_bulkio_input
for line in strip_comments(open(filename, 'r')):
File "/usr/local/redhawk/core/bin/update_project", line 86, in strip_comments
ch += safe_next(chars)
File "/usr/local/redhawk/core/bin/update_project", line 56, in safe_next
return next(item)
NameError: global name 'next' is not defined
I would appreciate suggestions on how to convert 1.8.4 device to 1.9 device.

Answer 1

Based on your error message "NameError: global name 'next' is not defined" and the contents of the 1.9.0 release of update_project python script, I am assuming that you are running a version of python less than 2.6. The next function is a python builtin that was introduced in Python 2.6 (http://docs.python.org/2/library/functions.html#next). This is a known bug in the upgrade script as it should be compatible with Python 2.4 as well as Python 2.6 (The default python installations in CentOS 5 and 6 respectively). To fix this, you may modify the update_project script located in $OSSIEHOME/bin/update_project and define the following function:
if not hasattr(__builtins__, 'next'):
# Python 2.4 does not have a free-standing next() function
def next(iterator, default):
"""
Backwards compatibility next() equivalent for Python 2.4.
"""
try:
return iterator.next()
except StopIteration:
return default
You should then remove the previously defined "safe_next" function.
Lastly, you need to replace the two calls to "safe_next" with a call to the newly implement next function and add a second argument of empty string ''
For clarity, a diff of update_project with these changes is below:
## -46,16 +46,16 ## Options:
_bulkio_re = re.compile('BULKIO_data[A-Za-z]+_In_i')
-def safe_next(item):
- """
- Returns the next value of the iterator, or an empty string if the end of
- iteration has been reached. Allows string processing to gracefully handle
- the end of a line without explicit catch statements.
- """
- try:
- return next(item)
- except StopIteration:
- return ''
+if not hasattr(__builtins__, 'next'):
+ # Python 2.4 does not have a free-standing next() function
+ def next(iterator, default):
+ """
+ Backwards compatibility next() equivalent for Python 2.4.
+ """
+ try:
+ return iterator.next()
+ except StopIteration:
+ return default
def strip_comments(source):
"""
## -75,7 +75,7 ## def strip_comments(source):
# Look for end comment token; if the end of the line is reached
# ch will be an empty string
while ch == '*':
- ch = safe_next(chars)
+ ch = next(chars, '')
if ch == '/':
inComment = False
break
## -83,7 +83,7 ## def strip_comments(source):
if ch == '/':
# Read the next character to see if it matches a comment token
# (if it does not, both characters will be added to the output)
- ch += safe_next(chars)
+ ch += next(chars, '')
if ch == '/*':
# Comment, start discarding
inComment = True