Re: [rdiff-backup-users] OverflowError reading get_sigs() response in very large tree

[Charles Duffy]

dean gaudet wrote:
> i wonder why i don't have this problem on million+ inode systems... i've 
> had it working with python2.3 before too (although i'm using 2.4 now)... 
> i'm still on 1.0.x.
>
> is there a 32/64-bit mismatch between your hosts?  any chance there's some 
> bug in that?
>   

Both systems are 64-bit.

Also, looking at the code for Python 2.3, I was wrong to specify 
PY_SSIZE_T_MAX; it's rather INT_MAX which is compared against (until 
Python 2.5, the first release using PY_SSIZE_T_MAX). I'll be building a 
debug version of Python 2.3 and looking at what's going on from gdb 
today as well as checking for whether the PY_SSIZE_T patch resolved the 
issue in Python 2.5.

INT_MAX is the same on both platforms (2147483647), and I'm inclined to 
call it a pretty safe bet we aren't actually transferring that much 
data. (I was looking at an even larger number at first because 
sys.maxint returns LONG_MAX, which is *very* different between the two 
platforms, and was speculating about issues relating to INT_MAX being 
unable to fit into an ssize_t; however, ssize_t appears to be an 
integer, like INT_MAX, so this suspicion appears unfounded).

I'll keep you (and the rest of the list) informed.