Opened 10 years ago
Closed 10 years ago
#2854 closed defect (fixed)
travis failing regress
Reported by: | robe | Owned by: | strk |
---|---|---|---|
Priority: | medium | Milestone: | PostGIS 2.2.0 |
Component: | QA/buildbots | Version: | 2.1.x |
Keywords: | Cc: |
Description
Travis has been failing regress since my #899 commits, but the errors don't seem to be breaking any where there so might be something unrelated.
Errors from travis
Error encountered loading /home/travis/build/postgis/postgis/regress/00-regress-install/share/contrib/postgis/postgis.sql, see /tmp/pgis_reg/regress_log for details make[1]: *** [check] Error 3 make[1]: Leaving directory `/home/travis/build/postgis/postgis/regress' make: *** [check] Error 1 The command "make test" exited with 2.
Change History (5)
comment:1 by , 10 years ago
Owner: | changed from | to
---|---|
Status: | new → assigned |
comment:2 by , 10 years ago
Here it is: 00-regress-install/lib/postgis-2.2.so: undefined symbol: HeapTupleHeaderGetDatum
From https://travis-ci.org/postgis/postgis/builds/31147383#L1944
Thats with PostgreSQL 9.1.4: https://travis-ci.org/postgis/postgis/builds/31147383#L962
Do Debbie or Winnie build against 9.1.4 ?
comment:3 by , 10 years ago
The compiler invocation has an explicit -I/usr/include, which comes before -I/usr/include/postgresql/9.1/server, so I'm thinking there might be a conflict between the postgresql header used and the library linked against.
comment:4 by , 10 years ago
debbie builds against PostgreSQL 9.1.9 winnie builds too against PostgreSQL 9.1.9 on x86_64-w64-mingw32 (testing in mingw mode instead of edb mode)
On last run both 9.1s passed.
comment:5 by , 10 years ago
Resolution: | → fixed |
---|---|
Status: | assigned → closed |
Was a problem about having mixed postgres installs, fixed as of r12845: https://travis-ci.org/postgis/postgis/builds/31227108
Error encountered loading /home/travis/build/postgis/postgis/regress/00-regress-install/share/contrib/postgis/postgis.sql, see /tmp/pgis_reg/regress_log for details
I'm fixing run_test to honour -v in this case, so we get the actual error message