Source: graphite2
Version: 1.1.3-1
Severity: wishlist
Tags: sid upstream
User: debian-powerpc...@breakpoint.cc
Usertags: powerpcspe
Hi,
graphite currently FTBFS[1] on powerpcspe[2]. This is due to a failing test
("25/70 Test #25: scher1Output"). The difference between the expected and
actual output is:
$ diff -u tests/standards/scher1.log build/tests/scher1.log
--- tests/standards/scher1.log 2012-05-30 10:17:49.000000000 +0000
+++ build/tests/scher1.log 2013-04-28 13:58:37.374119778 +0000
@@ -8,7 +8,7 @@
04 971 3@220,1035 3.9 0.1 1 30 5 5 654 654
05 965 4@100,1330 3.8 1.9 1 30 4 4 64e 64e
06 1233 3@150,250 0.4 0.0 1 30 6 6 627 627
-07 965 6@47,949 0.0 -0.4 1 30 7 7 64e 64e
+07 965 6@47,949 -0.0 -0.4 1 30 7 7 64e 64e
Advance width = 10.7
Char Unicode Before After
$
I.e. a rounding issue leading to "-0.0" instead of expected "0.0". This happens
in gr2fonttest/gr2FontTest.cpp:688 where an fprintf() prints a float via %6.1f
and actually has sth. like -0.00xxx in float orgX.
So I'm wondering if we are strictly looking for an orgX value of >0.00... here,
or if we accept a rounding tolerance of small values -0.1 < x < 0. In the
former case, we would need to fix powerpcspe float rounding somehow. In the
latter case, a more relaxed rounding tolerance in gr2fonttest/gr2FontTest.cpp
would be best.
As the maintainer (or upstream), you have a better understanding what's
appropriate here. Depending on your decision, I can try to provide a
patch.
(This issue also applies to graphite2_1.2.1-1 in experimental.)
Thanks in advance,
Roland
[1]
http://buildd.debian-ports.org/status/fetch.php?pkg=graphite2&arch=powerpcspe&ver=1.1.3-1&stamp=1361527940
[2] http://wiki.debian.org/PowerPCSPEPort
-- System Information:
Debian Release: 7.0
APT prefers unreleased
APT policy: (500, 'unreleased'), (500, 'unstable')
Architecture: powerpcspe (ppc)
Kernel: Linux 3.8.0 (SMP w/2 CPU cores)
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8) (ignored: LC_ALL
set to en_GB.UTF-8)
Shell: /bin/sh linked to /bin/dash
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
