Hi All, Was trying to build osmocom 3g system. Its failing in running test on osmo-hlr, any clues? ## --------------------------------- #### osmo-hlr 0.1.0.7-8db4 test suite. #### --------------------------------- ## Regression tests. 1: auc ok 2: auc_ts_55_205_test_sets ok 3: gsup_server ok 4: db FAILED (testsuite.at:30) ## ------------- #### Test results. #### ------------- ## ERROR: All 4 tests were run,1 failed unexpectedly.
On Thu, Nov 16, 2017 at 02:59:04PM +0000, Dees wrote:
Hi All, Was trying to build osmocom 3g system. Its failing in running test on osmo-hlr, any clues? ## --------------------------------- #### osmo-hlr 0.1.0.7-8db4 test suite. #### --------------------------------- ## Regression tests. 1: auc ok 2: auc_ts_55_205_test_sets ok 3: gsup_server ok 4: db FAILED (testsuite.at:30) ## ------------- #### Test results. #### ------------- ## ERROR: All 4 tests were run,1 failed unexpectedly.
Hi Dees,
the error from your log is:
-DDB (2067) abort at 18 in [INSERT INTO subscriber (imsi) VALUES ($imsi)]: UNIQUE constraint failed: subscriber.imsi +DDB (2067) abort at 35 in [INSERT INTO subscriber (imsi) VALUES ($imsi)]: UNIQUE constraint failed: subscriber.imsi
which is the identical error as the test run intends to produce, except that "at 18" became "at 35" for no apparent reason.
I wonder what that number is supposed to mean ... 35 could point at the IMSI value, while 18 doesn't make any sense at all.
I cannot reproduce your failure locally, my test outcome expects '18', and that's what the test gives me. It has been '18' from the start in the git history of that file.
Anyway, this is an error message produced by sqlite, and maybe they have fixed the number '18' to point closer to the IMSI instead?
I am using libsqlite 3.16.2-5 on Debian 9, what version are you running?
...we may need to disable that error output for the test or mangle the output somehow...
~N
Hi Neels, Thanks for checking. I was trying to build on Ubutun LTS 14.04 and sqlite3 version is 3.8.2. I believe 3.8.2 is latest stable version in 14.04. Kind regards,Dees
root@ip-172-31-7-26:/home/ubuntu# lsb_release -aNo LSB modules are available.Distributor ID: UbuntuDescription: Ubuntu 14.04.5 LTSRelease: 14.04Codename: trusty
root@ip-172-31-7-26:/home/ubuntu# sqlite3 -version3.8.2 2013-12-06 14:53:30 27392118af4c38c5203a04b8013e1afdb1cebd0d
From: Neels Hofmeyr nhofmeyr@sysmocom.de To: Dees motosingh@yahoo.co.uk Cc: "openbsc@lists.osmocom.org" openbsc@lists.osmocom.org Sent: Sunday, 19 November 2017, 13:10 Subject: Re: [osmo-hlr 0.1.0.7-8db4] testsuite: 4 failed
On Thu, Nov 16, 2017 at 02:59:04PM +0000, Dees wrote:
Hi All, Was trying to build osmocom 3g system. Its failing in running test on osmo-hlr, any clues? ## --------------------------------- #### osmo-hlr 0.1.0.7-8db4 test suite. #### --------------------------------- ## Regression tests. 1: auc ok 2: auc_ts_55_205_test_sets ok 3: gsup_server ok 4: db FAILED (testsuite.at:30) ## ------------- #### Test results. #### ------------- ## ERROR: All 4 tests were run,1 failed unexpectedly.
Hi Dees,
the error from your log is:
-DDB (2067) abort at 18 in [INSERT INTO subscriber (imsi) VALUES ($imsi)]: UNIQUE constraint failed: subscriber.imsi +DDB (2067) abort at 35 in [INSERT INTO subscriber (imsi) VALUES ($imsi)]: UNIQUE constraint failed: subscriber.imsi
which is the identical error as the test run intends to produce, except that "at 18" became "at 35" for no apparent reason.
I wonder what that number is supposed to mean ... 35 could point at the IMSI value, while 18 doesn't make any sense at all.
I cannot reproduce your failure locally, my test outcome expects '18', and that's what the test gives me. It has been '18' from the start in the git history of that file.
Anyway, this is an error message produced by sqlite, and maybe they have fixed the number '18' to point closer to the IMSI instead?
I am using libsqlite 3.16.2-5 on Debian 9, what version are you running?
...we may need to disable that error output for the test or mangle the output somehow...
~N
On Mon, Nov 20, 2017 at 03:36:51PM +0000, Dees wrote:
Hi Neels, Thanks for checking. I was trying to build on Ubutun LTS 14.04 and sqlite3 version is 3.8.2. I believe 3.8.2 is latest stable version in 14.04. Kind regards,Dees
14.04 is rather old, but that's not an excuse for having a regression test vulnerable for false positives. So I pushed https://gerrit.osmocom.org/4948 for review.
To work a round the issue until that is merged, you could either skip the 'make check' or update the expected output to match your current message:
cd osmo-hlr/tests/db make update_exp
~N
-
Dees -
Neels Hofmeyr