Eric Searcy wrote: > That diff doesn't look like trunk, in my working copy the line > `CHILD[fd] = 0' is after the block you are concerned with. Please > update your sources (I have r428 currently).
If this has changed then just say so. But I had the very latest svn at the time and this code was exactly as I showed in my diff. I've never touched popen.c except to add debugging to determine this problem. Really. > Yes, but 2.2.1 has this same block of code that you are concerned about > removing. It's at line 621 in popen.c, and its also an essential piece > of code that prevents waiting for pid==0. What I sent was a diff against 2.2.1. So the block was moved or something because diff isn't in the habit of lying. > latest version of trunk. The code was improved to prevent buffer > overflow errors in 415, which introduced a regression---a regression > which has since been fixed in trunk (from my July 20th email ``patch > [1/6] r415 fix''). I was testing with 425. Assuming you still have the error with r428, could you post the relevant > debug output? There is no relevant debug output. There's no logging in that function at all, which is why I added the Verbose line to confirm where it was happening. (I added a lot, then removed the others when I confirmed where it was happening) Anyway, it does seem to be working now. The fix between 425 and 428 did it, which is clearly -- Jo Rhett senior geek Silicon Valley Colocation _______________________________________________ Bug-cfengine mailing list [email protected] https://cfengine.org/mailman/listinfo/bug-cfengine
