Re: [Nmh-workers] Artificial argv limits.

[Paul Fox]

ralph wrote:
 > Hi Paul (Fox),
 > 
 > > but looking at rmm.c, refile.c, and folder_delmsgs(), i don't think
 > > that issue applies.  the patch below fixes the problem, and makes
 > > rmmproc and refile do the right thing (well, at least, the modern
 > > thing).  NB: i didn't test to the limit where execvp should return
 > > E2BIG.
 > 
 > When you say it makes refile `do the right thing', do you mean refile
 > doesn't create copies of the emails if the future rmmproc can't be
 > invoked?  I'm guessing not from the patch's brevity.  :-)  That means
 > the user is left with many duplicate emails, most likely with different
 > numbers, hard-linked across two folders to clean up.

as far as i can tell, this was also true before my change.  are you
suggesting otherwise?

(one way to reduce the likelihood of this, and of the original
overflow issue, might be to pass the message arguments to rmmproc in
the same form that they were received by refile.  currently "refile
all" causes a call to rmmproc with a full enumeration of message
numbers.  if "refile all" instead caused "rmmproc all", on the
assumption that the same expansion would result, then the original
invocation of refile would be as likely to fail as rmmproc.  there are
probably subtleties to this i'm not getting right now.)

paul
=---------------------
 paul fox, address@hidden (arlington, ma, where it's 31.8 degrees)