--- Begin Message ---
Package: release.debian.org
Severity: normal
User: release.debian.org@packages.debian.org
Usertags: unblock
Please unblock package dsdp
This upload fixes the RC bug #857067. I have attached a debdiff against
the version in stretch.
unblock dsdp/5.8-9.4
Regards,
James
diff -Nru dsdp-5.8/debian/changelog dsdp-5.8/debian/changelog
--- dsdp-5.8/debian/changelog 2012-05-27 21:01:57.000000000 +0100
+++ dsdp-5.8/debian/changelog 2017-03-28 18:22:35.000000000 +0100
@@ -1,3 +1,34 @@
+dsdp (5.8-9.4) unstable; urgency=medium
+
+ * Non-maintainer upload.
+ * Revert previous patches in 5.8-9.2 and 5.8-9.3, they are completely wrong
+ and end up causing *flags to always be 0 on 64-bit big-endian systems.
+ * Use correct integer type for Fortran prototypes and variables
+ (Closes: #857067)
+
+ -- James Clarke <jrtc27@debian.org> Tue, 28 Mar 2017 18:22:35 +0100
+
+dsdp (5.8-9.3) unstable; urgency=medium
+
+ * Non-maintainer upload.
+ * Initialize all INFO vars. Closes: #857067
+ * Revert unneeded changes in d/rules
+
+ -- Ole Streicher <olebole@debian.org> Tue, 28 Mar 2017 17:40:25 +0200
+
+dsdp (5.8-9.2) unstable; urgency=medium
+
+ * Non-maintainer upload.
+
+ [ Matthias Klose ]
+ * Build using -O2 on s390x.
+
+ [ Dimitri John Ledkov ]
+ * Cast INFO to int before storing it in the flag. LP: #1543982.
+ Closes: #857067
+
+ -- Andreas Tille <tille@debian.org> Tue, 28 Mar 2017 09:22:18 +0200
+
dsdp (5.8-9.1) unstable; urgency=low
* Non-maintainer upload.
diff -Nru dsdp-5.8/debian/patches/series dsdp-5.8/debian/patches/series
--- dsdp-5.8/debian/patches/series 1970-01-01 01:00:00.000000000 +0100
+++ dsdp-5.8/debian/patches/series 2017-03-28 18:22:35.000000000 +0100
@@ -0,0 +1 @@
+type-mismatch.patch
diff -Nru dsdp-5.8/debian/patches/type-mismatch.patch dsdp-5.8/debian/patches/type-mismatch.patch
--- dsdp-5.8/debian/patches/type-mismatch.patch 1970-01-01 01:00:00.000000000 +0100
+++ dsdp-5.8/debian/patches/type-mismatch.patch 2017-03-28 18:20:40.000000000 +0100
@@ -0,0 +1,39 @@
+Description: Use correct integer type for Fortran prototypes and variables
+ GNU Fortran's default integer width is 32-bit, the same as GCC, therefore use
+ int rather than long int when interfacing with Fortran. This was an issue on
+ 64-bit big-endian systems, since the upper 32 bits of the long would be set,
+ which would also be lost when truncating to a 32-bit integer.
+Author: James Clarke <jrtc27@debian.org>
+Bug-Debian: https://bugs.debian.org/857067
+Last-Update: 2017-03-28
+---
+This patch header follows DEP-3: http://dep.debian.net/deps/dep3/
+--- a/include/dsdplapack.h
++++ b/include/dsdplapack.h
+@@ -4,11 +4,11 @@
+ \file dsdplapack.h
+ \brief DSDP uses BLAS and LAPACK for many of its operations.
+ */
+-
+-typedef long int ffinteger;
+ /*
+-typedef int ffinteger;
++typedef long int ffinteger;
+ */
++typedef int ffinteger;
++
+ /*
+ #define __DSDP_NONAMEMANGLING
+ #undef __DSDP_NONAMEMANGLING
+--- a/src/vecmat/dtrsm2.c
++++ b/src/vecmat/dtrsm2.c
+@@ -1,7 +1,7 @@
+ #include "dsdplapack.h"
+
+-typedef long int integer;
+-typedef long int logical;
++typedef int integer;
++typedef int logical;
+
+ #define max(a,b) ((a) >= (b) ? (a) : (b))
+ #define dmax(a,b) (double)max(a,b)
--- End Message ---