Re: gettext-0.14.2 fails on parallel build

[Bruno Haible]

Alexandre Duret-Lutz wrote:
> a fix for this belongs to the rule that starts those multiple `$(MAKE)
> elc-stamp' processes, not to the elc-stamp rule.  IOW locks are
> not needed in regular builds, they would only be needed to
> prevent multiple *recover* rules to start `$(MAKE) elc-stamp'.

Ok, what about this (untested) patch?

diff -c -3 -r1.45 lisp.am
*** lib/am/lisp.am      16 Mar 2005 00:10:30 -0000      1.45
--- lib/am/lisp.am      17 Mar 2005 19:00:16 -0000
***************
*** 49,60 ****
  $(am__ELCFILES): elc-stamp
  ## Recover from the removal of address@hidden
  ##
! ## Make sure not to call `make elc-stamp' if emacs is not available,
! ## because as all *.elc files appear as missing, a parallel make would
! ## attempt to build elc-stamp several times.
        @if test "$(EMACS)" != no && test ! -f $@; then \
!         rm -f elc-stamp; \
!         $(MAKE) $(AM_MAKEFLAGS) elc-stamp; \
        else : ; fi
  
  
--- 49,72 ----
  $(am__ELCFILES): elc-stamp
  ## Recover from the removal of address@hidden
  ##
! ## Do not call `make elc-stamp' if emacs is not available, because it would
! ## be useless.
        @if test "$(EMACS)" != no && test ! -f $@; then \
! ## Mutual exclusion locking to ensure that if multiple processes need to
! ## rebuild elc-stamp, only the first one does it, and the followers wait.
!         trap 'rm -rf elc-lock elc-stamp' 1 2 3 15; \
!         if mkdir elc-lock 2>/dev/null; then \
! ## This code is being executed by the first process.
!           rm -f elc-stamp; \
!           $(MAKE) $(AM_MAKEFLAGS) elc-stamp; \
!           rmdir elc-lock; \
!         else \
! ## This code is being executed by the follower processes.
! ## Wait until the first process is done.
!           while test -d elc-lock; do sleep 1; done; \
! ## Succeed if and only if the first process succeeded.
!           test -f elc-stamp; exit $$?; \
!         fi; \
        else : ; fi