Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 23 May 2006 22:36:54 +0300
From:      Andrey Slusar <anrays@gmail.com>
To:        Mark Evenson <evenson@panix.com>, <m.boyarov@gmail.com>
Cc:        freebsd-ports@freebsd.org
Subject:   Re: setting EMACS_PORT_NAME=emacs22 forks make infinitely at the port registration step
Message-ID:  <86bqtoy1m1.fsf@santinel.home.ua>
In-Reply-To: <e4vm6k$obv$1@sea.gmane.org> (Mark Evenson's message of "Tue, 23 May 2006 21:01:38 %2B0200")
References:  <200204301328.g3UDS9X43078@thelonious.eziba.com> <e4vm6k$obv$1@sea.gmane.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Tue, 23 May 2006 21:01:38 +0200, Mark Evenson wrote:

> Jay Sachs wrote:
> >> Number:         37596
> >> Category:       ports
> >> Synopsis:       EMACS_PORT_NAME=xemacs21 forks make infinitely when installing other ports

 This bug is uncorrectly fixed.

> When setting EMACS_PORT_NAME=emacs22 (The new Emacs has some nice new
> features esp. wrt. Lisp development) on the command line or in
> /etc/make.conf, the ports mechanism forks infinitely at the port
> registration step. Eventually this forking stops when the recursion
> limit is reached, but gosh does it take a long time on a slow
> machine. Or even a fast machine for that matter.

> This seems to be identical to the symptom that was fixed with
> ports/37596, and I don't really understand why it would be
> returning. Maybe it is from some other cause?

> This is happening on freshly supped ports trees under RELENG_6_1 on
> several machines (two new, one quite old that has been nursed along
> over the years since 2001), using portupgrade -ras to keep ports in
> sync with dependencies.

> The recursively invoked make(1)s look like this from a ps:

> root     86542  8.4  0.4  1284  1136  p3  S+    8:54PM   0:00.28 make
> CHILD_DEPENDS=yes PARENT_CHECKED= package-depends-list

> The empty PARENT_CHECKED is suspicious, but when I run manually with a
> value, it still forks forever.

> Am I using the EMACS_PORT_NAME variable correctly?  (i.e. its not a
> forbidden user knob like USE_XXX)

> Can I help provide more information?

> Should I reopen the bug report or start a new one?

 See this fix, submitted by nobutaka. This is now in testing and later
committed to ports tree.
--8<---------------cut here---------------start------------->8---
--- Mk/bsd.port.mk.orig	Tue May 23 22:21:46 2006
+++ Mk/bsd.port.mk	Tue May 23 22:22:03 2006
@@ -1339,7 +1339,7 @@
 .include "${PORTSDIR}/Mk/bsd.openssl.mk"
 .endif
 
-.if defined(USE_EMACS) || defined(EMACS_PORT_NAME)
+.if defined(USE_EMACS)
 .include "${PORTSDIR}/Mk/bsd.emacs.mk"
 .endif
 
--8<---------------cut here---------------end--------------->8---

-- 
Regards,
Andrey.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?86bqtoy1m1.fsf>