Bug#127472: does not build from source on powerpc and s390
Package: libsafe
Version: 2.0-9-1
Severity: serious
Tags: patch
Hi,
libsafe doesn't build from source on s390 and powerpc. va_list assignment is
not portable. Please use the attached patch. For details of the error see
http://buildd.debian.org/build.php?arch=&pkg=libsafe--- libsafe-2.0-9/src/intercept.c Thu Nov 29 21:34:30 2001
+++ libsafe-2.0-9-s390/src/intercept.c Wed Jan 2 16:20:13 2002
@@ -576,7 +576,8 @@
/*
* We have to save a copy of ap, since va_arg() will modify it.
*/
- va_list orig_ap = ap;
+ va_list orig_ap;
+ __va_copy(orig_ap, ap);
for (; count>0; count--) {
p = va_arg(ap, char *);
@@ -585,7 +586,7 @@
}
}
- ap = orig_ap;
+ __va_copy(ap, orig_ap);
}
res = real_vfprintf(fp, format, ap);
@@ -645,7 +646,8 @@
/*
* We have to save a copy of ap, since va_arg() will modify it.
*/
- va_list orig_ap = ap;
+ va_list orig_ap;
+ __va_copy(orig_ap, ap);
for (; count>0; count--) {
p = va_arg(ap, char *);
@@ -654,7 +656,7 @@
}
}
- ap = orig_ap;
+ __va_copy(ap, orig_ap);
}
res = real_vfprintf(fp, format, ap);
Bug#127739: please add patch to support s390
Package: libhoard Version: 2.1.0-1 Severity: serious Tags: patch Hi, could you please apply the attached patch to support s390. I have sent it already to upstream since he contacted me regarding my previous NMU. Thanks, Gerharddiff -urN libhoard-2.1.0/arch-specific.cpp libhoard-2.1.0-s390/arch-specific.cpp --- libhoard-2.1.0/arch-specific.cpp Fri Jan 4 06:44:38 2002 +++ libhoard-2.1.0-s390/arch-specific.cpp Fri Jan 4 06:44:04 2002 @@ -293,7 +293,30 @@ : "r"(oldval), "r"(newval), "r"(*oldval) : "cr0", "memory"); #endif -#if !(defined(__sparc__) || defined(i386) || defined(__sgi) || defined(__powerpc__)) +#if defined(__s390__) +__asm__ __volatile__( + " lhi 1,3\n" /* CS must be aligned on 4 byte b. */ + " nr1,%1\n" /* isolate last 2 bits of address */ + " xr%1,1\n" /* make addr % 4 == 0 */ + " sll 1,3\n" + " ar%0,1\n" /* add alignement to bitnr */ + " lhi 1,31\n" + " nr1,%0\n" /* make shift value */ + " xr%0,1\n" + " srl %0,3\n" + " la%1,0(%0,%1)\n" /* calc. address for CS */ + " lhi 2,1\n" + " sll 2,0(1)\n" /* make OR mask */ + " l %0,0(%1)\n" + "0: lr1,%0\n" /* CS loop starts here */ + " xr1,2\n" /* set bit */ + " cs%0,1,0(%1)\n" + " jl0b\n" + " nr%0,2\n" /* isolate old bit */ + : "+a" (newval), "+a" (oldval) : + : "cc", "memory", "1", "2" ); +#endif +#if !(defined(__sparc__) || defined(i386) || defined(__sgi) || defined(__powerpc__) || !defined(__s390__)) #error "Hoard does not include support for user-level locks for this platform." #endif return newval;
Bug#130077: please apply attached patch to support s390
Package: clisp Version: 2.27-0.2 Severity: serious Tags: patch Hi, could you please apply the attached patch to support s390. It consists mainly of the ffcall patch which I have already sent to the ffcall Debian maintainer and the upstream author. I have also disabled global register assignment on s390 for the lisp stack address which doesn't seem to work on s390. It builds then fine. Since clisp seems to be orphaned I will do a source NMU if nobody minds. Thanks, Gerhard clisp-s390.diff.gz Description: GNU Zip compressed data
Bug#160057: wrong assumption on char signedness
Package: dstooltk Version: 2.0-5 Severity: important Hi, char is per default unsigned on arm, powerpc and s390. There is a warning load_procs.c:152: warning: comparison is always true due to limited range of data type during build of your package that will cause a runtime error. Please check http://buildd.debian.org/fetch.php?&pkg=dstooltk&ver=2.0-5&arch=s390&stamp=1031345044&file=log&as=raw for details. Thanks, Gerhard
Bug#159098: FTBFS: autoheader fails
Package: aethera Version: 0.9.3-8 Severity: important Hi, during a recent rebuild of your package with gcc-3.2 on s390 the build failed. This was reproducible with gcc-2.95. For more details see http://people.debian.org/~gt/gcc-3.2_transition/failed/common/aethera_0.9.3-8.0.1_20020821-0904 Thanks Gerhard
Bug#166738: FTBFS with gcc 3.2
Package: korelib Version: 0.0.1-9 Severity: important Hi, the package get at least the following compile errors c++ -I. -I. -I.. -I../include -g -O2 -Wall -Wp,-MD,.deps/kernel.pp -c -fPIC -DPIC kernel.cpp -o .libs/kernel.lo In file included from kernel.cpp:4: ../include/kore/servicemanager.h:199: `hash' was not declared in this scope ../include/kore/servicemanager.h:199: parse error before `char' ../include/kore/servicemanager.h:212: 'srv_hash_type' is used as a type, but is not defined as a type. Thanks, Gerhard
Bug#177289: wrong assumption on char signedness
Package: ubit Version: 2.11.0-2 Severity: important Hi, char is per default unsigned on arm, powerpc and s390. There are several warnings, e.g. ums.cc:834: warning: comparison is always true due to limited range of data type during build of your package that might cause a runtime error. Please check http://buildd.debian.org/fetch.php?&pkg=ubit&ver=2.11.0-2&arch=s390&stamp=1038922818&file=log&as=raw for details. Thanks, Gerhard
Bug#177431: wrong assumption on char signedness
Package: siag Version: 3.5.3-1 Severity: important Hi, char is per default unsigned on arm, powerpc and s390. There are several warnings, e.g. cmds.c:2516: warning: comparison is always false due to limited range of data type during build of your package that might cause a runtime error. Please check http://buildd.debian.org/fetch.php?&pkg=siag&ver=3.5.3-1&arch=s390&stamp=1041104785&file=log&as=raw for details. Thanks, Gerhard
Bug#177436: wrong assumption on char signedness
Package: sted2 Version: 2.07m+20010303-6 Severity: important Hi, char is per default unsigned on arm, powerpc and s390. There are several warnings, e.g. assign.c:231: warning: comparison is always true due to limited range of data type during build of your package that might cause a runtime error. Please check http://buildd.debian.org/fetch.php?&pkg=sted2&ver=2.07m%2B20010303-6&arch=s390&stamp=1041274887&file=log&as=raw for details. Thanks, Gerhard
Bug#182157: wrong assumption on char signedness
Package: blackened Version: 1.8.1-3 Severity: important Hi, char is per default unsigned on arm, powerpc and s390. There iare two warnings fileio.c:648: warning: comparison is always true due to limited range of data type fileio.c:697: warning: comparison is always true due to limited range of data type during build of your package that might cause a runtime error. Please check http://buildd.debian.org/fetch.php?&pkg=blackened&ver=1.8.1-3&arch=s390&stamp=1044806788&file=log&as=raw for details. In your case the variable c should be defined as int. Defining it as char is even wrong on i386, since the char 0xff is then interpreted as EOF. On the platforms mentioned above the EOF condition is never true, leading to an infinite loop. Thanks, Gerhard
Bug#185299: Please upgrade to latest config.sub and config.guess
Package: sabre Version: 0.2.4b-16 Severity: serious Hi, the package doesn't build on s390 since config.sub and config.guess are not up-to-date. Could you upgrade them please. The package autotools-dev provides them in /usr/share/misc. Regards, Gerhard
