Re: gcc/f/intdoc.in

[Jeffrey A Law <law at upchuck dot cygnus dot com>]

  In message <19990215055330.29112.qmail@deer>you write:
  > I have my "cvs" directory in which I do this stuff.  I've done "cvs checkou
  > t
  > egcs" to get the mainline, "cvs checkout -r egcs_1_1_branch egcs-1.1" to
  > get the branch.  Then I've done "cvs update" on each, on occasion.
  > 
  > A few patches I've created (usually on copies of these directories,
  > so I can isolate what I'm doing and review it, after which I use
  > `diff' to create a patch which I then apply to the cvs/ directories)
  > have been applicable to both checked-out directories.
  > 
  > So, after applying the patches to both directories, I've done:
  > 
  >   cvs commit -m whatever egcs egcs-1.1
  > 
  > And, the response I've has strongly suggested that both branches' files
  > have been updated accordingly, just as if I'd done:
OK.  I think this is OK.

  > Two things I've noticed: one, the single-command approach yields
  > a curious/spurious diagnostic after it's seemingly done; two,
  > I don't get an email from egcs-cvs that notifies me (us) about the
  > changes to the egcs_1_1_branch branch.
Odd.

  > Checking in egcs-1.1/gcc/f/version.c;
  > /egcs/carton/cvsfiles/egcs/gcc/f/version.c,v  <--  version.c
  > new revision: 1.13.2.6; previous revision: 1.13.2.5
  > done
  > Cannot open file /tmp/#egcscvs.lastdir.6986.
  > [craig@deer cvs]$
[ ... ]
Ah.  Yes.  This is why we're not getting checkin messages.

I'll see if I can get something similar to happen with a multi-directory
checkin.  Though I must admit the mail wrappers have always been a little
bit of a black box to me.

Jeff