Re: [GNU PSPP 0.8.1.1-g608cc6] testsuite: 144 failed

[Ben Pfaff]

This is odd, but it reminds me of commit 4afd608d61b469 "Avoid redundant
linking against libtool convenience libraries."  So: can you try this
patch?

diff --git a/utilities/automake.mk b/utilities/automake.mk
index 5a44ff8..d667d62 100644
--- a/utilities/automake.mk
+++ b/utilities/automake.mk
@@ -12,5 +12,4 @@ utilities_pspp_convert_SOURCES = utilities/pspp-convert.c
 utilities_pspp_convert_CPPFLAGS = $(AM_CPPFLAGS) -DINSTALLDIR=\"$(bindir)\"
 utilities_pspp_convert_LDADD = \
        src/libpspp/liblibpspp.la \
-       src/libpspp-core.la \
-       gl/libgl.la
+       src/libpspp-core.la


On Sun, Jan 05, 2014 at 03:42:36PM -0500, Jeremy Lavergne wrote:
> Alrighty, I?m a bit rusty on getting into the test cases for PSPP but I think 
> I got it: hmap_first_with_hash(const struct hmap *map, size_t hash) isn?t 
> happy.
> 

> $ sudo 
> DYLD_LIBRARY_PATH="/opt/pspp/var/macports/build/_Users_aeetes_dports_math_pspp-devel/pspp-devel/work/pspp-0.8.1.1-g608cc6/src/.libs:$DYLD_LIBRARY_PATH"
>  lldb ../../../utilities/.libs/pspp-convert  hotel.csv
> 2014-01-05 15:39:59.503 lldb[601:212f] Metadata.framework [Error]: couldn't 
> get the client port
> Current executable set to '../../../utilities/.libs/pspp-convert' (x86_64).
> (lldb) run
> Process 602 launched: '../../../utilities/.libs/pspp-convert' (x86_64)
> Process 602 stopped
> * thread #1: tid = 0x1977, 0x00000001001555e8 
> libpspp-core-0.8.1.1-g608cc6.dylib`create_iconv__ [inlined] 
> hmap_first_with_hash(member_offset=0) + 10 at hmap.h:308, queue = 
> 'com.apple.main-thread, stop reason = EXC_BAD_ACCESS (code=1, address=0x0)
>     frame #0: 0x00000001001555e8 
> libpspp-core-0.8.1.1-g608cc6.dylib`create_iconv__ [inlined] 
> hmap_first_with_hash(member_offset=0) + 10 at hmap.h:308
>    305        static inline struct hmap_node *
>    306        hmap_first_with_hash (const struct hmap *map, size_t hash)
>    307        {
> -> 308          return hmap_find_hash__ (map->buckets[hash & map->mask], 
> hash);
>    309        }
>    310        
>    311        /* Returns the next node in MAP after NODE that has the same 
> hash
> (lldb) bt
> * thread #1: tid = 0x1977, 0x00000001001555e8 
> libpspp-core-0.8.1.1-g608cc6.dylib`create_iconv__ [inlined] 
> hmap_first_with_hash(member_offset=0) + 10 at hmap.h:308, queue = 
> 'com.apple.main-thread, stop reason = EXC_BAD_ACCESS (code=1, address=0x0)
>     frame #0: 0x00000001001555e8 
> libpspp-core-0.8.1.1-g608cc6.dylib`create_iconv__ [inlined] 
> hmap_first_with_hash(member_offset=0) + 10 at hmap.h:308
>     frame #1: 0x00000001001555de 
> libpspp-core-0.8.1.1-g608cc6.dylib`create_iconv__ [inlined] 
> hmapx_first_with_hash at hmapx.h:296
>     frame #2: 0x00000001001555de 
> libpspp-core-0.8.1.1-g608cc6.dylib`create_iconv__(tocode=0x000000010018b2fe, 
> fromcode=0x000000010050dbd0) + 59 at i18n.c:72
>     frame #3: 0x0000000100154699 
> libpspp-core-0.8.1.1-g608cc6.dylib`recode_substring_pool [inlined] 
> create_iconv(tocode=0x000000010018b2fe, fromcode=0x000000010050dbd0) + 11 at 
> i18n.c:92
>     frame #4: 0x000000010015468e 
> libpspp-core-0.8.1.1-g608cc6.dylib`recode_substring_pool(to=<unavailable>, 
> from=<unavailable>, pool=0x0000000100803820, text=<unavailable>) + 56 at 
> i18n.c:525
>     frame #5: 0x000000010014525b 
> libpspp-core-0.8.1.1-g608cc6.dylib`sfm_open_reader [inlined] 
> parse_header(header=0x00000000324c4624, r=0x000000010050d790, 
> info=0x000000010050d620, dict=0x000000010050f380) + 35 at 
> sys-file-reader.c:1039
>     frame #6: 0x0000000100145238 
> libpspp-core-0.8.1.1-g608cc6.dylib`sfm_open_reader(fh=<unavailable>, 
> encoding=0x0000000000000000, dictp=0x00007fff5fbff6f8, 
> infop=0x0000000000000000) + 7030 at sys-file-reader.c:527
>     frame #7: 0x0000000100003fd7 pspp-convert`main(argc=<unavailable>, 
> argv=<unavailable>) + 907 at pspp-convert.c:159
>     frame #8: 0x00007fff9559c5fd libdyld.dylib`start + 1
>     frame #9: 0x00007fff9559c5fd libdyld.dylib`start + 1

> 
> 
> On Dec 29, 2013, at 13:52, Ben Pfaff <address@hidden> wrote:
> 
> > On Sat, Dec 28, 2013 at 04:56:54PM -0500, Jeremy Lavergne wrote:
> >> Test 144:
> >> test-source: line 69: 94163 Segmentation fault: 11  pspp-convert hotel.sav 
> >> hotel.csv
> >> 
> >> If it?s new, let me know if you?d like in to debug it further. I
> >> believe gdb is available, however I?m pretty sure LLDB is the program
> >> du jour for clang debugging.
> > 
> > This is new, and I can't seem to reproduce it on my machine.  I'm happy
> > with a backtrace from gdb or lldb, either is fine.
> > 
> > Thanks,
> > 
> > Ben.
>