Bug 34459

From: "Thomas Hurst" <freaky@aagh.net>
>         When configure fails, bsd.port.mk dumps config.log to the
>         terminal.  This typcially results in several thousand lines of
>         crap which often hides the actual problem way up in the display
>         buffer, assuming it extends that far up.
>
>         It should either pipe it through a pager, tail the last 20 lines
>         or so (since that'll be where it failed anyway), or be made
>         optional.

bsd.port.mk doesn't dump the contents of the config.log.  config.log is
dumped by the configure script in the work/<distname> directory.

Scot

* Scot W. Hetzel (hetzels@westbend.net) wrote:

> From: "Thomas Hurst" <freaky@aagh.net>
> >         When configure fails, bsd.port.mk dumps config.log to the
> >         terminal.  This typcially results in several thousand lines of
> >         crap which often hides the actual problem way up in the display
> >         buffer, assuming it extends that far up.
> >
> >         It should either pipe it through a pager, tail the last 20 lines
> >         or so (since that'll be where it failed anyway), or be made
> >         optional.
> 
> bsd.port.mk doesn't dump the contents of the config.log.  config.log is
> dumped by the configure script in the work/<distname> directory.

Um, line 2270:

${CONFIGURE_ENV} ./${CONFIGURE_SCRIPT} ${CONFIGURE_ARGS}; then \
        ${ECHO} "===>  Script \"${CONFIGURE_SCRIPT}\" failed: here are the contents of \"${CONFIGURE_LOG}\""; \
        ${CAT} ${CONFIGURE_LOG}; \
        ${ECHO} "(end of \"${CONFIGURE_LOG}\")"; \

-- 
Thomas 'Freaky' Hurst  -  freaky@aagh.net  -  http://www.aagh.net/
-
Binary, adj.:
	Possessing the ability to have friends of both sexes.

This is nasty. Somebody please patch this out, or at least give it a knob
that defaults to 'no'. I'm sorry to be so blunt, but dumping config.log in
its entirety sucks really hard.

-- 
Alan E
"Please rush me my portable walrus cleaning kit! Yes I am over 18, 
but my IQ isn't."

From: "Thomas Hurst" <freaky@aagh.net>
> To: "Scot W. Hetzel" <hetzels@westbend.net>
>  > bsd.port.mk doesn't dump the contents of the config.log.  config.log is
>  > dumped by the configure script in the work/<distname> directory.
>
>  Um, line 2270:
>
>  ${CONFIGURE_ENV} ./${CONFIGURE_SCRIPT} ${CONFIGURE_ARGS}; then \
>          ${ECHO} "===>  Script \"${CONFIGURE_SCRIPT}\" failed: here are
the contents of \"${CONFIGURE_LOG}\""; \
>          ${CAT} ${CONFIGURE_LOG}; \
>          ${ECHO} "(end of \"${CONFIGURE_LOG}\")"; \
>
I stand corrected, I have attached a patch that puts a 'if CONFIG_DEBUG....
.. fi' around this.  I believe this was added so that bento's logs would
have the output of the config.log file in them.

Bento's scripts would need to be changed to add a -DCONFIG_DEBUG when
building packages.

Scot

Attached is a uuencoded version of the patch.


begin 666 bsd.port.mk.diff
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$9&EF"@``
`
end

Ignore the previous patches, as they were missing a "}" to CONFIG_DEBUG_TMP.

I have also added a code that will change the auto[conf|make] programs in the
configure script to the correct versions.  This is needed as I have noticed
that sometimes both autoconf* are run (i.e ${AUTOCONF} by bsd.port.mk, and
autoconf by the src distibutions Makefile).

Scot

Index: bsd.port.mk
===================================================================
RCS file: /home/ncvs/ports/Mk/bsd.port.mk,v
retrieving revision 1.397
diff -u -r1.397 bsd.port.mk
--- bsd.port.mk	24 Jan 2002 01:15:03 -0000	1.397
+++ bsd.port.mk	31 Jan 2002 16:32:34 -0000
@@ -513,6 +513,7 @@
 #				  to turn them off (default: ${ECHO}).
 # PATCH_DEBUG	- If set, print out more information about the patches as
 #				  it attempts to apply them.
+# CONFIG_DEBUG	- If set, print out more information about the configure process.
 # PKG_DBDIR		- Where package installation is recorded (default: /var/db/pkg)
 # NO_PKG_REGISTER - Don't register a port installation as a package.
 # FORCE_PKG_REGISTER - If set, it will overwrite any existing package
@@ -1112,6 +1113,12 @@
 PATCH_DIST_ARGS+=	-C
 .endif
 
+.if defined(CONFIG_DEBUG)
+CONFIG_DEBUG_TMP=	yes
+.else
+CONFIG_DEBUG_TMP=	no
+.endif
+
 .if exists(/bin/tar)
 TAR?=	/bin/tar
 .else
@@ -2197,6 +2204,9 @@
 	fi
 .if defined(HAS_CONFIGURE)
 	@(cd ${CONFIGURE_WRKSRC} && \
+		${PERL} -pi -e  "s/aclocal/${ACLOCAL}/g; s/automake/${AUTOMAKE}/g; \
+				 s/autoconf/${AUTOCONF}/g; s/autoheader/${AUTOHEADER}/g;" \
+				${CONFIGURE_SCRIPT} ; \
 		if ! ${SETENV} CC="${CC}" CXX="${CXX}" \
 	    CFLAGS="${CFLAGS}" CXXFLAGS="${CXXFLAGS}" \
 	    INSTALL="/usr/bin/install -c -o ${BINOWN} -g ${BINGRP}" \
@@ -2204,9 +2214,13 @@
 	    INSTALL_PROGRAM="${INSTALL_PROGRAM}" \
 	    INSTALL_SCRIPT="${INSTALL_SCRIPT}" \
 	    ${CONFIGURE_ENV} ./${CONFIGURE_SCRIPT} ${CONFIGURE_ARGS}; then \
-			${ECHO} "===>  Script \"${CONFIGURE_SCRIPT}\" failed: here are the contents of \"${CONFIGURE_LOG}\""; \
-			${CAT} ${CONFIGURE_LOG}; \
-			${ECHO} "(end of \"${CONFIGURE_LOG}\")"; \
+			if [ ${CONFIG_DEBUG_TMP} = yes ]; then \
+				${ECHO} "===>  Script \"${CONFIGURE_SCRIPT}\" failed: here are the contents of \"${CONFIGURE_LOG}\""; \
+				${CAT} ${CONFIGURE_LOG}; \
+				${ECHO} "(end of \"${CONFIGURE_LOG}\")"; \
+			else \
+				${ECHO} "===> Script  \"${CONFIGURE_SCRIPT}\" failed: review ${CONFIGURE_WRKSRC}/${CONFIGURE_LOG} for problem" ; \
+			fi ; \
 			${FALSE}; \
 		fi)
 .endif

Comment 7 Pete Fritchman 2002-02-20 08:07:59 UTC

Responsible Changed
From-To: freebsd-ports->portmgr

Over to maintainer.  A fix in the meantime is to set CONFIGURE_LOG=/dev/null 
in /etc/make.conf.

Comment 8 Maxim Sobolev 2002-04-25 16:31:21 UTC

State Changed
From-To: open->closed

The problem was resolved in rev.1.410 of bsd.port.mk. Thank you for submission!