As per summary, this is going to replace dev-ruby/ruby-postgres so please keyword it, then we can ask for it to go stable for all of you at once (rather than splitting it down). Thanks, Diego
I've been a bit too trigger-happy, sorry.
prefix done, thx
Already done ~hppa in bug #322175.
~x86-fbsd done.
@ alpha: please keyword >=dev-ruby/pg-0.14.1 and up This does not depend on rspec-1 anymore.
Created attachment 373960 [details] pg build log with failing test phase
dev-ruby/pg-0.14.1 fails tests: >>> Source compiled. * Running test phase for ruby19 ... Setting up test database for PG_Connection tests Running initdb Creating the test DB .................................*....................*....Tearing down test database No pidfile (#<Pathname:/var/tmp/portage/dev-ruby/pg-0.14.1/work/ruby19/pg-0.14.1/tmp_test_specs/data/postmaster.pid>) Setting up test database for PG_Result tests Creating the test DB ..F.....F.................Tearing down test database .F Pending: PG::Connection discards previous results (if any) before waiting on an #async_exec # Not yet implemented # ./spec/pg/connection_spec.rb:521 PG::Connection multinationalization support rubyforge #22925: m17n support works around the unsupported JOHAB encoding by returning stuff in 'ASCII_8BIT' # figuring out how to create a string in the JOHAB encoding # ./spec/pg/connection_spec.rb:801 Failures: 1) PG::Result encapsulates errors in a PGError object Failure/Error: result.error_field( PG::PG_DIAG_SOURCE_LINE ).should == '857' expected: "857" got: "984" (using ==) # ./spec/pg/result_spec.rb:74:in `block (2 levels) in <top (required)>' 2) PG::Result should return the parameter type of the specified prepared statement parameter Failure/Error: @conn.prepare( 'queryfinder', query ) PG::Error: ERROR: column "current_query" does not exist LINE 1: ...* FROM pg_stat_activity WHERE user = $1::name AND current_qu... ^ # ./spec/pg/result_spec.rb:132:in `prepare' # ./spec/pg/result_spec.rb:132:in `block (2 levels) in <top (required)>' 3) PG knows whether or not the library is threadsafe Failure/Error: PG.should be_threadsafe() expected threadsafe? to return true, got false # ./spec/pg_spec.rb:27:in `block (2 levels) in <top (required)>' Finished in 30.42 seconds 87 examples, 3 failures, 2 pending Failed examples: rspec ./spec/pg/result_spec.rb:52 # PG::Result encapsulates errors in a PGError object rspec ./spec/pg/result_spec.rb:130 # PG::Result should return the parameter type of the specified prepared statement parameter rspec ./spec/pg_spec.rb:26 # PG knows whether or not the library is threadsafe * ERROR: dev-ruby/pg-0.14.1::gentoo failed (test phase): I'll check the newer versions in a moment.
Both 0.17.0-r1 and 0.17.1 have this test fail: Failures: 1) PG knows whether or not the library is threadsafe Failure/Error: PG.should be_threadsafe() expected threadsafe? to return true, got false # ./spec/pg_spec.rb:27:in `block (2 levels) in <top (required)>'
Thanks for the quick test. The failure for pg-0.14.1 looks related to bug #457704
(In reply to Tobias Klausmann from comment #8) > Both 0.17.0-r1 and 0.17.1 have this test fail: > > Failures: > > 1) PG knows whether or not the library is threadsafe > Failure/Error: PG.should be_threadsafe() > expected threadsafe? to return true, got false > # ./spec/pg_spec.rb:27:in `block (2 levels) in <top (required)>' Perhaps you have USE=-threads for postgresql-base?
(In reply to Hans de Graaff from comment #10) > > 1) PG knows whether or not the library is threadsafe > > Failure/Error: PG.should be_threadsafe() > > expected threadsafe? to return true, got false > > # ./spec/pg_spec.rb:27:in `block (2 levels) in <top (required)>' > > Perhaps you have USE=-threads for postgresql-base? Yes, I can reproduce the problem with USE=-threads. I've added an additional test dependency on postgresql-base[threads] to fix this.
After a mere four years, I get to close this bug... Keyworded =dev-ruby/pg-0.17.1 on alpha.