Sorry to send another message like this, but..
3706 works, 3718 fails.
Looks like the migration process is a little odd after that point.
On 3718:
C:\rails\spending_plan>rake test:units
(in C:/rails/spending_plan)
-- create_table("activity_codes", {:force=>true})
-> 0.2850s
-- create_table("budget_entities", {:force=>true})
-> 0.1110s
-- create_table("category_codes", {:force=>true})
-> 0.1270s
-- create_table("l2_codes", {:force=>true})
-> 0.2370s
-- create_table("plan_elements", {:force=>true})
rake aborted!
OCIError: ORA-00907: missing right parenthesis: CREATE TABLE plan_elements (id N
UMBER(38) NOT NULL PRIMARY KEY, spending_plan_id NUMBER(38), category_code_id NU
MBER(38), approved_budget_amount NUMBER(9), expenditures_amount NUMBER(9), proje
cted_amount NUMBER(9) DEFAULT 0, amendment_amount NUMBER(9) DEFAULT 0, lock_vers
ion NUMBER(38) DEFAULT 0, updated_at DATE(0), created_at DATE(0), comment_text V
ARCHAR2(255), proposed_amount NUMBER(9) DEFAULT 0, activity_code_id NUMBER(38))
On 3706:
C:\rails\spending_plan>rake test:units
(in C:/rails/spending_plan)
-- create_table("activity_codes", {:force=>true})
-> 0.1900s
-- create_table("budget_entities", {:force=>true})
-> 0.1420s
-- create_table("category_codes", {:force=>true})
-> 0.0950s
-- create_table("l2_codes", {:force=>true})
-> 0.0950s
-- create_table("plan_elements", {:force=>true})
-> 0.1110s
-- create_table("spending_plans", {:force=>true})
-> 0.5860s
-- initialize_schema_information()
-> 0.0000s
-- columns("schema_info")
-> 0.0470s
c:/ruby/bin/ruby -Ilib;test
"c:/ruby/lib/ruby/gems/1.8/gems/rake-0.7.0/lib/rake/
rake_test_loader.rb" "test/unit/activity_code_test.rb"
"test/unit/budget_entity_
test.rb" "test/unit/category_code_test.rb"
"test/unit/l2_code_test.rb" "test/uni
t/plan_element_test.rb" "test/unit/spending_plan_test.rb"
"test/unit/user_test.r
b"
Loaded suite c:/ruby/lib/ruby/gems/1.8/gems/rake-0.7.0/lib/rake/rake_test_loader
Started
.................
Finished in 1.218 seconds.
17 tests, 116 assertions, 0 failures, 0 errors
Wilson Bilkovich wrote:> Sorry to send another message like this, but.. > 3706 works, 3718 fails. > Looks like the migration process is a little odd after that point.I''ll take a look.
Michael Schoen wrote:> Wilson Bilkovich wrote: >> Sorry to send another message like this, but.. >> 3706 works, 3718 fails. >> Looks like the migration process is a little odd after that point. > > I''ll take a look.Fixed as part of http://dev.rubyonrails.org/ticket/4230
On 3/14/06, Michael Schoen <schoenm@earthlink.net> wrote:> Michael Schoen wrote: > > Wilson Bilkovich wrote: > >> Sorry to send another message like this, but.. > >> 3706 works, 3718 fails. > >> Looks like the migration process is a little odd after that point. > > > > I''ll take a look. > > Fixed as part of > > http://dev.rubyonrails.org/ticket/4230 >Thanks again for your work on Oracle support. I get a lot of mileage out of it.
> Thanks again for your work on Oracle support. I get a lot of mileage out of it.Sorry for the troubles, I''ll be sure to submit my patches to this list before making widespread changes like that :) -- Rick Olson http://techno-weenie.net
On 3/14/06, Michael Schoen <schoenm@earthlink.net> wrote:> > Michael Schoen wrote: > > Wilson Bilkovich wrote: > >> Sorry to send another message like this, but.. > >> 3706 works, 3718 fails. > >> Looks like the migration process is a little odd after that point. > > > > I''ll take a look. > > Fixed as part of > > http://dev.rubyonrails.org/ticket/4230 > > > Should AR pass all tests on oracle? I''m on rev 3900 atm and I still getfour errors. -- Corey Donohoe http://www.atmos.org/ corey.donohoe@gmail.com _______________________________________________ Rails-core mailing list Rails-core@lists.rubyonrails.org http://lists.rubyonrails.org/mailman/listinfo/rails-core
> Should AR pass all tests on oracle? I''m on rev 3900 atm and I still get > four errors. > > > -- > Corey Donohoe > http://www.atmos.org/ > corey.donohoe@gmail.comCorey showed me the errors http://rafb.net/paste/results/5O41hK24.html (will self destruct in 24 hours. #1-3 look like fixture/constraint issues. But #4 is the result of an alias that''s too long for oracle. join_category_nonexistent_posts Any thoughts? Should Oracle have a special aliasing method that converts that to j_c_posts? Should all databases do that consistently? I''d really like to have readable aliases and get away from something like j_c_posts_2. -- Rick Olson http://techno-weenie.net
Older versions of DB2 suffer the same character length issue with table names as an FYI. On 3/16/06, Rick Olson <technoweenie@gmail.com> wrote:> > > Should AR pass all tests on oracle? I''m on rev 3900 atm and I still get > > four errors. > > > > > > -- > > Corey Donohoe > > http://www.atmos.org/ > > corey.donohoe@gmail.com > > Corey showed me the errors http://rafb.net/paste/results/5O41hK24.html > (will self destruct in 24 hours. #1-3 look like fixture/constraint > issues. But #4 is the result of an alias that''s too long for oracle. > join_category_nonexistent_posts > > Any thoughts? > > Should Oracle have a special aliasing method that converts that to > j_c_posts? Should all databases do that consistently? I''d really > like to have readable aliases and get away from something like > j_c_posts_2. > > -- > Rick Olson > http://techno-weenie.net > _______________________________________________ > Rails-core mailing list > Rails-core@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/rails-core >_______________________________________________ Rails-core mailing list Rails-core@lists.rubyonrails.org http://lists.rubyonrails.org/mailman/listinfo/rails-core
Rick Olson wrote:> Should Oracle have a special aliasing method that converts that to > j_c_posts? Should all databases do that consistently? I''d really > like to have readable aliases and get away from something like > j_c_posts_2.Do other dbs have a length limit on aliases? Oracle''s is 30 chars.
Rick Olson wrote:> Should Oracle have a special aliasing method that converts that to > j_c_posts? Should all databases do that consistently? I''d really > like to have readable aliases and get away from something like > j_c_posts_2.Do other dbs have a length limit on aliases? Oracle''s is 30 chars.
DB2 version 7 and before limit for table/column/alias is 18 chars. Version 8 forwards is 30 for column/alias and 128 for tablenames. Not sure why IBM did it that way... On 3/16/06, Michael Schoen <schoenm@earthlink.net> wrote:> > Rick Olson wrote: > > Should Oracle have a special aliasing method that converts that to > > j_c_posts? Should all databases do that consistently? I''d really > > like to have readable aliases and get away from something like > > j_c_posts_2. > > Do other dbs have a length limit on aliases? Oracle''s is 30 chars. > > _______________________________________________ > Rails-core mailing list > Rails-core@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/rails-core >_______________________________________________ Rails-core mailing list Rails-core@lists.rubyonrails.org http://lists.rubyonrails.org/mailman/listinfo/rails-core
On Mar 16, 2006, at 1:38 PM, Steve Longdo wrote:> Older versions of DB2 suffer the same character length issue with > table names as an FYI.FYI, it looks like Sybase 12.5 is limited to 30 chars for both table and column names. John -- John R. Sheets http://umberdog.com/
On 3/16/06, John Sheets <fisheye@metacasa.net> wrote:> On Mar 16, 2006, at 1:38 PM, Steve Longdo wrote: > > > Older versions of DB2 suffer the same character length issue with > > table names as an FYI. > > FYI, it looks like Sybase 12.5 is limited to 30 chars for both table > and column names. > > John63 for psql, and 255 for mysql. You guys need to upgrade :) Seriously though, I''m adding a #table_alias_for method to the adapters to handle this stuff. I''ll add 30 for oracle and sybase. What''s the deal with DB2 though? Should I use 128? Are there a lot of DB2 v7 users out there? -- Rick Olson http://techno-weenie.net
This sounds great from the oracle side, can test for ya whenever. On 3/17/06, Rick Olson <technoweenie@gmail.com> wrote:> On 3/16/06, John Sheets <fisheye@metacasa.net> wrote: > > On Mar 16, 2006, at 1:38 PM, Steve Longdo wrote: > > > > > Older versions of DB2 suffer the same character length issue with > > > table names as an FYI. > > > > FYI, it looks like Sybase 12.5 is limited to 30 chars for both table > > and column names. > > > > John > > 63 for psql, and 255 for mysql. > > You guys need to upgrade :) > > Seriously though, I''m adding a #table_alias_for method to the adapters > to handle this stuff. I''ll add 30 for oracle and sybase. What''s the > deal with DB2 though? Should I use 128? Are there a lot of DB2 v7 > users out there? > > -- > Rick Olson > http://techno-weenie.net > _______________________________________________ > Rails-core mailing list > Rails-core@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/rails-core >-- Corey Donohoe http://www.atmos.org/ corey.donohoe@gmail.com
On 3/18/06, Corey Donohoe <corey.donohoe@gmail.com> wrote:> This sounds great from the oracle side, can test for ya whenever.I did it last night. Let me know how it works! I added column alias limits to the database adapters I knew about, and also reworked the way they''re generated. I put the table name (or STI name) in front. If it''s a dupe, I add the reflection name it came from, and then index numbers (_2, _3, etc). Join tables get a _join suffix instead of a prefix. The reasoning behind this is I didn''t want an aliased table name like foo_bar_baz_posts to shorten to foo_bar_ba. Instead I''d rather have posts_foo_bar_baz shorten to posts_foo_b. -- Rick Olson http://techno-weenie.net
3921 fixed the identifier too long problem. thanks dude. I get different results when I run all.sh than when I run a specific test, anyone have any idea why that might be? On 3/18/06, Rick Olson <technoweenie@gmail.com> wrote:> On 3/18/06, Corey Donohoe <corey.donohoe@gmail.com> wrote: > > This sounds great from the oracle side, can test for ya whenever. > > I did it last night. Let me know how it works! > > I added column alias limits to the database adapters I knew about, and > also reworked the way they''re generated. I put the table name (or STI > name) in front. If it''s a dupe, I add the reflection name it came > from, and then index numbers (_2, _3, etc). Join tables get a _join > suffix instead of a prefix. The reasoning behind this is I didn''t > want an aliased table name like foo_bar_baz_posts to shorten to > foo_bar_ba. Instead I''d rather have posts_foo_bar_baz shorten to > posts_foo_b. > > -- > Rick Olson > http://techno-weenie.net > _______________________________________________ > Rails-core mailing list > Rails-core@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/rails-core >-- Corey Donohoe http://www.atmos.org/ corey.donohoe@gmail.com