[Debian-astro-commits] [eso-midas] 01/01: Workaround problem in gfortran-6 optimization. Closes: #831134
Ole Streicher
olebole at moszumanska.debian.org
Mon Aug 1 09:31:21 UTC 2016
This is an automated email from the git hooks/post-receive script.
olebole pushed a commit to branch debian
in repository eso-midas.
commit 6451e6a7ba4c88d0ea83dbf0e0f5833fdca1c0e6
Author: Ole Streicher <olebole at debian.org>
Date: Mon Aug 1 11:21:06 2016 +0200
Workaround problem in gfortran-6 optimization. Closes: #831134
---
debian/changelog | 9 +++++----
debian/control | 2 +-
debian/patches/fix_madrid.patch | 25 +++++++++++++++++++++++++
debian/patches/series | 1 +
4 files changed, 32 insertions(+), 5 deletions(-)
diff --git a/debian/changelog b/debian/changelog
index a9ee636..2e9aa5e 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,9 +1,10 @@
-eso-midas (15.09pl1.0-2) UNRELEASED; urgency=low
+eso-midas (15.09pl1.0-2) unstable; urgency=medium
- * Update Standards-Version to 3.9.7 (no changes)
- * Add ASCL-Id, update VCS fields
+ * Update Standards-Version to 3.9.8 (update VCS fields)
+ * Add ASCL-Id
+ * Workaround problem in gfortran-6 optimization. Closes: #831134
- -- Ole Streicher <olebole at debian.org> Thu, 03 Mar 2016 17:21:08 +0100
+ -- Ole Streicher <olebole at debian.org> Mon, 01 Aug 2016 11:18:30 +0200
eso-midas (15.09pl1.0-1) unstable; urgency=low
diff --git a/debian/control b/debian/control
index 426a6e4..89a66da 100644
--- a/debian/control
+++ b/debian/control
@@ -11,7 +11,7 @@ Build-Depends: debhelper (>= 9),
libreadline-dev,
libx11-dev,
libxt-dev
-Standards-Version: 3.9.7
+Standards-Version: 3.9.8
Homepage: http://www.eso.org/projects/esomidas/
Vcs-Git: https://anonscm.debian.org/git/debian-astro/packages/eso-midas.git
Vcs-Browser: https://anonscm.debian.org/cgit/debian-astro/packages/eso-midas.git
diff --git a/debian/patches/fix_madrid.patch b/debian/patches/fix_madrid.patch
new file mode 100644
index 0000000..cd3d8bf
--- /dev/null
+++ b/debian/patches/fix_madrid.patch
@@ -0,0 +1,25 @@
+Author: Ole Streicher <olebole at debian.org>
+Description: Fix size of MADRID array in test
+ When gfortran-6 is called with an optimization (-O1 or higher), arrays with a
+ length 1 seem to be converted to a scalar and assumed to always return the same
+ value.
+ .
+ This fails with the MADRID array, which is a placeholder for an array.
+ Specifically ftoc_commf.for fails then, which leads to a missing ftoc_comm.h.
+ .
+ The simple workaround is to use a size > 1 here.
+ .
+ Note that the MADRID array is defined with a length=1 in ~350 places all over
+ eso-midas. They are not changed yet since the verification succeeds. If we get
+ strange bugreports, this may be considered as the problem.
+--- a/system/ftoc-new/ftoc_commf.for
++++ b/system/ftoc-new/ftoc_commf.for
+@@ -34,7 +34,7 @@
+ C
+ PROGRAM F2CCOM
+ C
+- INTEGER MADRID(1)
++ INTEGER MADRID(3)
+ INTEGER VALUE(3)
+ INTEGER STAT,MANY,NOPIX,NOBYT
+ INTEGER*8 PNTR
diff --git a/debian/patches/series b/debian/patches/series
index daae03f..ebacf71 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -14,3 +14,4 @@ fix_rarthm_for.patch
fix_mdb_put.patch
fix_fitswdb.patch
fix_selechar_c.patch
+fix_madrid.patch
--
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-astro/packages/eso-midas.git
More information about the Debian-astro-commits
mailing list