How do I get a ticket out of the ''unverified'' state? I have a patch for a new database adapter that I submitted over 2 months ago (http://dev.rubyonrails.org/ticket/4093) that has made no progress. It passes all of the unit tests. I have to keep updating my patch every couple of days as the edge moves forward. What is my next step in order to get my patch checked into the main source tree? I''d be willing to help someone with commit access get FrontBase setup in their environment so that they could run the tests themselves as an independent verification. I considered making it a plug-in, but I won''t be able to run the ActiveRecord unit tests if I go that route and it will be more difficult to keep it compatible as Rails evolves. Thanks, Mike
On Thu, Apr 20, 2006 at 08:33:13PM -0400, Mike Laster wrote:> How do I get a ticket out of the ''unverified'' state? I have a patch > for a new database adapter that I submitted over 2 months ago > (http://dev.rubyonrails.org/ticket/4093) that has made no progress. > > It passes all of the unit tests. I have to keep updating my patch > every couple of days as the edge moves forward. What is my next step > in order to get my patch checked into the main source tree? I''d be > willing to help someone with commit access get FrontBase setup in > their environment so that they could run the tests themselves as an > independent verification. > > I considered making it a plug-in, but I won''t be able to run the > ActiveRecord unit tests if I go that route and it will be more > difficult to keep it compatible as Rails evolves.Hey Mike. Thanks for staying on top of the adapter so that it remains up to date. Apologies that we haven''t gotten around to it sooner. It''s quite frustrating to put work into something only to see it sit around. I installed Frontbase and the ruby bindings for it. At first when I ran the test it was complaining that FBSQL_Connect.connect was getting one too many arguments passed to it. Just to move forward I took off the extra argument. Then when I run the tests I get loads of warnings from Object#type on line 511 of the frontbase adapter. if pk && klass.columns_hash[pk].type == :integer I am assuming klass.columns_hash[pk] is evaluating to nil. Aside from all the warnings, this is the result of running the tests: 853 tests, 375 assertions, 63 failures, 708 errors This is the primary error: ActiveRecord::StatementInvalid: FBError: Semantic error 231. INSERT value doesn''t match column: bonus_time. Semantic error 485. Near: INSERT INTO "topics"("author_name","title","approved","id","replies_count","bonus_time","content","written_on","author_email_address","last_read") VALUES(''David'',''The First Topic'',FALSE,1,0,TIMESTAMP ''2005-01-30 08:28:00'',''Have a nice day'',TIMESTAMP ''2003-07-16 09:28:00'',''david@loudthinking.com'',DATE ''2004-04-15'');. Semantic error 485. Near: DATE ''2004-04-15''. It could entirely be a result of me more or less blindly removing that extra parameter or not setting everything up correctly. Perhaps you may have a sense of what may be wrong here, or perhaps our neglect has gotten your latest patch out of date though since it was posted today that seems unlikely. Please advise :) On a side note, there is a bit of foo != nil or foo == nil in the code. The preferred approach in those cases is to use the nil? predicate method. In other words !foo.nil? or foo.nil?. marcel -- Marcel Molina Jr. <marcel@vernix.org>
On 4/20/06, Marcel Molina Jr. <marcel@vernix.org> wrote:> > Hey Mike. Thanks for staying on top of the adapter so that it remains up to > date. Apologies that we haven''t gotten around to it sooner. It''s quite > frustrating to put work into something only to see it sit around.It has been a bit, but it also a learning experience. :-)> I installed Frontbase and the ruby bindings for it. At first when I ran the > test it was complaining that FBSQL_Connect.connect was getting one too many > arguments passed to it. Just to move forward I took off the extra argument. > Then when I run the tests I get loads of warnings from Object#type on line > 511 of the frontbase adapter.Are you running the newest FrontBase? I am testing against 4.2.4. It definitely will not work with the 3.x line. About the extra argument...check that you have the newest Ruby bindings. The ones on the Frontbase site are *not* the newest. I had to modify them to work with ActiveRecord. I sent my changes back to them but they haven''t posted them yet. I have the source to the modified bindings checked in at: svn://rubyforge.org//var/svn/frontbase-rails/trunk/ruby-frontbase I added an extra argument to allow naming the database session, so it sounds like you are running the old 0.5.2 bindings. Once I get the adapter code checked in, I''ll look into releasing a gem of it to make things simpler. As of right now (revision 4241) it is passing the tests cleanly: 859 tests, 2866 assertions, 0 failures, 0 errors> On a side note, there is a bit of foo != nil or foo == nil in the code. The > preferred approach in those cases is to use the nil? predicate method. In > other words !foo.nil? or foo.nil?.I''ll be glad to change them. I was learning Ruby in the process of doing this so I''m sure some of it is still lacking in style :)
On Thu, Apr 20, 2006 at 11:28:24PM -0400, Mike Laster wrote:> On 4/20/06, Marcel Molina Jr. <marcel@vernix.org> wrote: > > I installed Frontbase and the ruby bindings for it. At first when I ran the > > test it was complaining that FBSQL_Connect.connect was getting one too many > > arguments passed to it. Just to move forward I took off the extra argument. > > Then when I run the tests I get loads of warnings from Object#type on line > > 511 of the frontbase adapter. > > Are you running the newest FrontBase? I am testing against 4.2.4. It > definitely > will not work with the 3.x line.I''ve got 4.2.4, yeah.> About the extra argument...check that you have the newest Ruby bindings. > The ones on the Frontbase site are *not* the newest. I had to modify > them to work > with ActiveRecord. I sent my changes back to them but they haven''t > posted them yet. > > I have the source to the modified bindings checked in at: > > svn://rubyforge.org//var/svn/frontbase-rails/trunk/ruby-frontbaseI did indeed have the bindings from the Frontbase site. When I grab yours from rubyforge, this is what I get when I run make: /usr/bin/ld: /Library/FrontBase/lib/libFBCAccess.a(FBCAccessAll-i386.o) malformed object (section (__TEXT,__textcoal_nt) no symbol at start of coalesced section) collect2: ld returned 1 exit status make: *** [frontbase.bundle] Error 1 (MacBook Pro)> > On a side note, there is a bit of foo != nil or foo == nil in the code. The > > preferred approach in those cases is to use the nil? predicate method. In > > other words !foo.nil? or foo.nil?. > > I''ll be glad to change them. I was learning Ruby in the process of doing this > so I''m sure some of it is still lacking in style :)No sweat. We all get better every day (sometimes ;)). marcel -- Marcel Molina Jr. <marcel@vernix.org>
On 4/20/06, Marcel Molina Jr. <marcel@vernix.org> wrote:> On Thu, Apr 20, 2006 at 11:28:24PM -0400, Mike Laster wrote: > > On 4/20/06, Marcel Molina Jr. <marcel@vernix.org> wrote: > > > I installed Frontbase and the ruby bindings for it. At first when I ran the > > > test it was complaining that FBSQL_Connect.connect was getting one too many > > > arguments passed to it. Just to move forward I took off the extra argument. > > > Then when I run the tests I get loads of warnings from Object#type on line > > > 511 of the frontbase adapter. > > > > Are you running the newest FrontBase? I am testing against 4.2.4. It > > definitely > > will not work with the 3.x line. > > I''ve got 4.2.4, yeah. > > > About the extra argument...check that you have the newest Ruby bindings. > > The ones on the Frontbase site are *not* the newest. I had to modify > > them to work > > with ActiveRecord. I sent my changes back to them but they haven''t > > posted them yet. > > > > I have the source to the modified bindings checked in at: > > > > svn://rubyforge.org//var/svn/frontbase-rails/trunk/ruby-frontbase > > I did indeed have the bindings from the Frontbase site. > > When I grab yours from rubyforge, this is what I get when I run make: > > /usr/bin/ld: /Library/FrontBase/lib/libFBCAccess.a(FBCAccessAll-i386.o) malformed object (section (__TEXT,__textcoal_nt) no symbol at start of coalesced section) > collect2: ld returned 1 exit status > make: *** [frontbase.bundle] Error 1 > > (MacBook Pro)Hmm...that is a new one to me. I don''t have access to an Intel Mac so I''ve only done this on PPC. That looks like FrontBase may have a bad framework in their distribution. I''ll forward this to their support people and see what they can find out.