head	1.1;
access;
symbols
	pkgsrc-2026Q1:1.1.0.44
	pkgsrc-2026Q1-base:1.1
	pkgsrc-2025Q4:1.1.0.42
	pkgsrc-2025Q4-base:1.1
	pkgsrc-2025Q3:1.1.0.40
	pkgsrc-2025Q3-base:1.1
	pkgsrc-2025Q2:1.1.0.38
	pkgsrc-2025Q2-base:1.1
	pkgsrc-2025Q1:1.1.0.36
	pkgsrc-2025Q1-base:1.1
	pkgsrc-2024Q4:1.1.0.34
	pkgsrc-2024Q4-base:1.1
	pkgsrc-2024Q3:1.1.0.32
	pkgsrc-2024Q3-base:1.1
	pkgsrc-2024Q2:1.1.0.30
	pkgsrc-2024Q2-base:1.1
	pkgsrc-2024Q1:1.1.0.28
	pkgsrc-2024Q1-base:1.1
	pkgsrc-2023Q4:1.1.0.26
	pkgsrc-2023Q4-base:1.1
	pkgsrc-2023Q3:1.1.0.24
	pkgsrc-2023Q3-base:1.1
	pkgsrc-2023Q2:1.1.0.22
	pkgsrc-2023Q2-base:1.1
	pkgsrc-2023Q1:1.1.0.20
	pkgsrc-2023Q1-base:1.1
	pkgsrc-2022Q4:1.1.0.18
	pkgsrc-2022Q4-base:1.1
	pkgsrc-2022Q3:1.1.0.16
	pkgsrc-2022Q3-base:1.1
	pkgsrc-2022Q2:1.1.0.14
	pkgsrc-2022Q2-base:1.1
	pkgsrc-2022Q1:1.1.0.12
	pkgsrc-2022Q1-base:1.1
	pkgsrc-2021Q4:1.1.0.10
	pkgsrc-2021Q4-base:1.1
	pkgsrc-2021Q3:1.1.0.8
	pkgsrc-2021Q3-base:1.1
	pkgsrc-2021Q2:1.1.0.6
	pkgsrc-2021Q2-base:1.1
	pkgsrc-2021Q1:1.1.0.4
	pkgsrc-2021Q1-base:1.1
	pkgsrc-2020Q4:1.1.0.2
	pkgsrc-2020Q4-base:1.1;
locks; strict;
comment	@# @;


1.1
date	2020.11.29.03.11.54;	author dholland;	state Exp;
branches;
next	;
commitid	4JvqUUOe7dlZjJxC;


desc
@@


1.1
log
@Make games/craft build again.

The custom C++ preprocessor thingy has been choking for a while on
clang; recently gcc's jointed the party. Work around this by building
it with -g instead of -O2. (For me at least with gcc, while -O2
produces a tool that gives spurious syntax errors, plain -O gives one
that loops forever instead, which isn't an improvement.)

Maybe someday someone will bother to find out where the undefined
behavior is.

Also build the actual game with -Wno-return-type to silence complaints
about dead functions not declared as dead and similar, and while here
adjust things so it still installs correctly regardless of umask.
@
text
@$NetBSD$

Build the preprocessor front end thingy with -g instead of with -O2.

With -O2 it crashes with a message about EOF inside a string constant;
this has been happening with clang for a while and gcc recently joined
the party. With just -O, it loops forever instead for me, which isn't
an improvement.

This should serve as a workaround until someone gets around to
figuring out where the undefined behavior is.

--- Makefile~	2020-11-29 02:36:49.745300963 +0000
+++ Makefile
@@@@ -42,7 +42,7 @@@@ depend:
 	@@false
 
 ${HCC}: hcc/hcc.cc
-	${CXX} ${CFLAGS} hcc/hcc.cc -o hcc/hcc
+	${CXX} ${CFLAGS:S/-O2/-g/} hcc/hcc.cc -o hcc/hcc
 
 install:
 	[ -d ${DESTDIR} ]
@