In my project, I am using 3 files throughout the whole process. The source file (.ada), a "three address code" file (.TAC), and my own temporary file for use during processing (.TACTMP).
in Caller.py:
TACFILE = open(str(sys.argv[1])[:-4] + ".TAC", 'w') # line 17
# Does a bunch of stuff
TACFILE.close() # line 653
# the below function is imported from Called.py
post_process_file_handler() # line 654
in Called.py:
TAC_FILE_NAME = str(sys.argv[1])[:-4] # line 6
TAC_lines = open(TAC_FILE_NAME + ".TAC", 'r').readlines() # line 7
If I try to run my program without already having a (even if it's blank) .TAC file, I will get the following error:
Traceback (most recent call last):
File "Caller.py", line 8, in <module>
from Called import post_process_file_handler
File "Called.py", line 7, in <module>
TAC_lines = file(TAC_FILE_NAME + ".TAC", 'r').readlines()
IOError: [Errno 2] No such file or directory: 'test76.TAC'
Why would this be happening? This error is being thrown even if I put a breakpoint at the beginning of Caller.py, well before the post_process_file_handler() function ever gets called.
For clarity: test76.TAC should be being generated by Caller.py, and then Called.py should open that file to process it further, for some reason that isn't happening.
This may be specific to my case, but I found out the issue is due to the order and manner in which I was using these streams.
In short, when the import line was encountered:
from Called import post_process_file_handler
it triggered some sort of initialization, and since the file pointer was a global variable in Called.py, it was initialized before Caller.py had a chance to create the .TAC file it would read from.
Moving the import line to just before I use the function fixed my issue, as nothing in Called.py is initialized until after Caller.py is done doing its work.