Rails core - Jul 2007 - Wrong interpretation of limit in both Mysql and PostgreSQL adapters

Can anyone check this ticket?
http://dev.rubyonrails.org/ticket/5489

It''s 1 year old already, but still valid imho.

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core-unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---

> Can anyone check this ticket?
> http://dev.rubyonrails.org/ticket/5489
>
> It''s 1 year old already, but still valid imho.
I don''t quite follow what the problem is with the current
implementation?  could you give me an example of what bugs /
mis-behaviour this patch will prevent?


-- 
Cheers

Koz

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core-unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---

Sure.

Well, I was asked to include #5484 into this ticket, so it is dealing
with distinct bugs at once.

To understand those bugs, let''s begin with a single migration that
creates a table with one column:
    t.column :limit3, :integer, :limit => 3

1) When you run this migration under postgresql, the column is dumped
to schema.rb without the limit:
    t.column "limit3", :integer

The patch in this ticket fixes the postgresql adapter to dump columns
with limits.

2) Also in postgresql, the above migration will generate a smallint
column, which uses 2 bytes for storage. I think it should generate an
integer column, so it can handle 3 bytes numbers without overflow --
and this is also covered by the patch attached.

3) Finally, in mysql the migration will generate a INT(3) column, what
means a column with display size of 3, and not a column with 3 bytes
for storage. The patch changes mysql adapter to use the limit
constraint to generate smallint, mediumint, int and bigint columns
accordingly.


On 7/3/07, Michael Koziarski <michael@koziarski.com>
wrote:
>
> > Can anyone check this ticket?
> > http://dev.rubyonrails.org/ticket/5489
> >
> > It''s 1 year old already, but still valid imho.
>
> I don''t quite follow what the problem is with the current
> implementation?  could you give me an example of what bugs /
> mis-behaviour this patch will prevent?
>
>
> --
> Cheers
>
> Koz
>
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core-unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---

> Well, I was asked to include #5484 into this ticket, so it is dealing
> with distinct bugs at once.
Sounds good.   This patch breaks sqlite''s tests,  if you can merge
those in, it seems good to apply.

> To understand those bugs, let''s begin with a single migration that
> creates a table with one column:
>     t.column :limit3, :integer, :limit => 3
>
> 1) When you run this migration under postgresql, the column is dumped
> to schema.rb without the limit:
>     t.column "limit3", :integer
>
> The patch in this ticket fixes the postgresql adapter to dump columns
> with limits.
>
> 2) Also in postgresql, the above migration will generate a smallint
> column, which uses 2 bytes for storage. I think it should generate an
> integer column, so it can handle 3 bytes numbers without overflow --
> and this is also covered by the patch attached.
>
> 3) Finally, in mysql the migration will generate a INT(3) column, what
> means a column with display size of 3, and not a column with 3 bytes
> for storage. The patch changes mysql adapter to use the limit
> constraint to generate smallint, mediumint, int and bigint columns
> accordingly.
>
>
> On 7/3/07, Michael Koziarski <michael@koziarski.com> wrote:
> >
> > > Can anyone check this ticket?
> > > http://dev.rubyonrails.org/ticket/5489
> > >
> > > It''s 1 year old already, but still valid imho.
> >
> > I don''t quite follow what the problem is with the current
> > implementation?  could you give me an example of what bugs /
> > mis-behaviour this patch will prevent?
> >
> >
> > --
> > Cheers
> >
> > Koz
> >
> > >
> >
>
> >
>

-- 
Cheers

Koz

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core-unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---

ops, I''m back.

well, sqlite doesn''t support :limit anyways, so I guess we should skip
this test for adapters that doesn''t support it, correct?

any ideas?

On 7/6/07, Michael Koziarski <michael@koziarski.com>
wrote:
>
> > Well, I was asked to include #5484 into this ticket, so it is dealing
> > with distinct bugs at once.
>
> Sounds good.   This patch breaks sqlite''s tests,  if you can merge
> those in, it seems good to apply.
>
>
> > To understand those bugs, let''s begin with a single migration
that
> > creates a table with one column:
> >     t.column :limit3, :integer, :limit => 3
> >
> > 1) When you run this migration under postgresql, the column is dumped
> > to schema.rb without the limit:
> >     t.column "limit3", :integer
> >
> > The patch in this ticket fixes the postgresql adapter to dump columns
> > with limits.
> >
> > 2) Also in postgresql, the above migration will generate a smallint
> > column, which uses 2 bytes for storage. I think it should generate an
> > integer column, so it can handle 3 bytes numbers without overflow --
> > and this is also covered by the patch attached.
> >
> > 3) Finally, in mysql the migration will generate a INT(3) column, what
> > means a column with display size of 3, and not a column with 3 bytes
> > for storage. The patch changes mysql adapter to use the limit
> > constraint to generate smallint, mediumint, int and bigint columns
> > accordingly.
> >
> >
> > On 7/3/07, Michael Koziarski <michael@koziarski.com> wrote:
> > >
> > > > Can anyone check this ticket?
> > > > http://dev.rubyonrails.org/ticket/5489
> > > >
> > > > It''s 1 year old already, but still valid imho.
> > >
> > > I don''t quite follow what the problem is with the
current
> > > implementation?  could you give me an example of what bugs /
> > > mis-behaviour this patch will prevent?
> > >
> > >
> > > --
> > > Cheers
> > >
> > > Koz
> > >
> > > >
> > >
> >
> > >
> >
>
>
> --
> Cheers
>
> Koz
>
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core-unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---

humm just reminded about this

tarmo added sqlite tests. can anyone check and hopefully add?

On Aug 1, 2007 5:39 PM, Pedro Mariano Belo <pedrobelo@gmail.com>
wrote:
> ops, I''m back.
>
> well, sqlite doesn''t support :limit anyways, so I guess we should
skip
> this test for adapters that doesn''t support it, correct?
>
> any ideas?
>
>
> On 7/6/07, Michael Koziarski <michael@koziarski.com> wrote:
> >
> > > Well, I was asked to include #5484 into this ticket, so it is
dealing
> > > with distinct bugs at once.
> >
> > Sounds good.   This patch breaks sqlite''s tests,  if you can
merge
> > those in, it seems good to apply.
> >
> >
> > > To understand those bugs, let''s begin with a single
migration that
> > > creates a table with one column:
> > >     t.column :limit3, :integer, :limit => 3
> > >
> > > 1) When you run this migration under postgresql, the column is
dumped
> > > to schema.rb without the limit:
> > >     t.column "limit3", :integer
> > >
> > > The patch in this ticket fixes the postgresql adapter to dump
columns
> > > with limits.
> > >
> > > 2) Also in postgresql, the above migration will generate a
smallint
> > > column, which uses 2 bytes for storage. I think it should
generate an
> > > integer column, so it can handle 3 bytes numbers without overflow
--
> > > and this is also covered by the patch attached.
> > >
> > > 3) Finally, in mysql the migration will generate a INT(3) column,
what
> > > means a column with display size of 3, and not a column with 3
bytes
> > > for storage. The patch changes mysql adapter to use the limit
> > > constraint to generate smallint, mediumint, int and bigint
columns
> > > accordingly.
> > >
> > >
> > > On 7/3/07, Michael Koziarski <michael@koziarski.com> wrote:
> > > >
> > > > > Can anyone check this ticket?
> > > > > http://dev.rubyonrails.org/ticket/5489
> > > > >
> > > > > It''s 1 year old already, but still valid imho.
> > > >
> > > > I don''t quite follow what the problem is with the
current
> > > > implementation?  could you give me an example of what bugs /
> > > > mis-behaviour this patch will prevent?
> > > >
> > > >
> > > > --
> > > > Cheers
> > > >
> > > > Koz
> > > >
> > > > >
> > > >
> > >
> > > >
> > >
> >
> >
> > --
> > Cheers
> >
> > Koz
> >
> > > >
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core-unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---

gee, no more sql files for creating test database structures?
i''ll try to update the patch once again

On Dec 3, 2007 5:52 PM, Pedro Mariano Belo <pedrobelo@gmail.com>
wrote:
> humm just reminded about this
>
> tarmo added sqlite tests. can anyone check and hopefully add?
>
>
> On Aug 1, 2007 5:39 PM, Pedro Mariano Belo <pedrobelo@gmail.com>
wrote:
> > ops, I''m back.
> >
> > well, sqlite doesn''t support :limit anyways, so I guess we
should skip
> > this test for adapters that doesn''t support it, correct?
> >
> > any ideas?
> >
> >
> > On 7/6/07, Michael Koziarski <michael@koziarski.com> wrote:
> > >
> > > > Well, I was asked to include #5484 into this ticket, so it
is dealing
> > > > with distinct bugs at once.
> > >
> > > Sounds good.   This patch breaks sqlite''s tests,  if you
can merge
> > > those in, it seems good to apply.
> > >
> > >
> > > > To understand those bugs, let''s begin with a single
migration that
> > > > creates a table with one column:
> > > >     t.column :limit3, :integer, :limit => 3
> > > >
> > > > 1) When you run this migration under postgresql, the column
is dumped
> > > > to schema.rb without the limit:
> > > >     t.column "limit3", :integer
> > > >
> > > > The patch in this ticket fixes the postgresql adapter to
dump columns
> > > > with limits.
> > > >
> > > > 2) Also in postgresql, the above migration will generate a
smallint
> > > > column, which uses 2 bytes for storage. I think it should
generate an
> > > > integer column, so it can handle 3 bytes numbers without
overflow --
> > > > and this is also covered by the patch attached.
> > > >
> > > > 3) Finally, in mysql the migration will generate a INT(3)
column, what
> > > > means a column with display size of 3, and not a column with
3 bytes
> > > > for storage. The patch changes mysql adapter to use the
limit
> > > > constraint to generate smallint, mediumint, int and bigint
columns
> > > > accordingly.
> > > >
> > > >
> > > > On 7/3/07, Michael Koziarski <michael@koziarski.com>
wrote:
> > > > >
> > > > > > Can anyone check this ticket?
> > > > > > http://dev.rubyonrails.org/ticket/5489
> > > > > >
> > > > > > It''s 1 year old already, but still valid
imho.
> > > > >
> > > > > I don''t quite follow what the problem is with
the current
> > > > > implementation?  could you give me an example of what
bugs /
> > > > > mis-behaviour this patch will prevent?
> > > > >
> > > > >
> > > > > --
> > > > > Cheers
> > > > >
> > > > > Koz
> > > > >
> > > > > >
> > > > >
> > > >
> > > > >
> > > >
> > >
> > >
> > > --
> > > Cheers
> > >
> > > Koz
> > >
> > > > > >
> > >
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core-unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---

ok, patch updated against r8279
can anyone verify?

On Dec 3, 2007 6:31 PM, Pedro Mariano Belo <pedrobelo@gmail.com>
wrote:
> gee, no more sql files for creating test database structures?
> i''ll try to update the patch once again
>
>
> On Dec 3, 2007 5:52 PM, Pedro Mariano Belo <pedrobelo@gmail.com>
wrote:
> > humm just reminded about this
> >
> > tarmo added sqlite tests. can anyone check and hopefully add?
> >
> >
> > On Aug 1, 2007 5:39 PM, Pedro Mariano Belo <pedrobelo@gmail.com>
wrote:
> > > ops, I''m back.
> > >
> > > well, sqlite doesn''t support :limit anyways, so I guess
we should skip
> > > this test for adapters that doesn''t support it, correct?
> > >
> > > any ideas?
> > >
> > >
> > > On 7/6/07, Michael Koziarski <michael@koziarski.com> wrote:
> > > >
> > > > > Well, I was asked to include #5484 into this ticket, so
it is dealing
> > > > > with distinct bugs at once.
> > > >
> > > > Sounds good.   This patch breaks sqlite''s tests, 
if you can merge
> > > > those in, it seems good to apply.
> > > >
> > > >
> > > > > To understand those bugs, let''s begin with a
single migration that
> > > > > creates a table with one column:
> > > > >     t.column :limit3, :integer, :limit => 3
> > > > >
> > > > > 1) When you run this migration under postgresql, the
column is dumped
> > > > > to schema.rb without the limit:
> > > > >     t.column "limit3", :integer
> > > > >
> > > > > The patch in this ticket fixes the postgresql adapter
to dump columns
> > > > > with limits.
> > > > >
> > > > > 2) Also in postgresql, the above migration will
generate a smallint
> > > > > column, which uses 2 bytes for storage. I think it
should generate an
> > > > > integer column, so it can handle 3 bytes numbers
without overflow --
> > > > > and this is also covered by the patch attached.
> > > > >
> > > > > 3) Finally, in mysql the migration will generate a
INT(3) column, what
> > > > > means a column with display size of 3, and not a column
with 3 bytes
> > > > > for storage. The patch changes mysql adapter to use the
limit
> > > > > constraint to generate smallint, mediumint, int and
bigint columns
> > > > > accordingly.
> > > > >
> > > > >
> > > > > On 7/3/07, Michael Koziarski
<michael@koziarski.com> wrote:
> > > > > >
> > > > > > > Can anyone check this ticket?
> > > > > > > http://dev.rubyonrails.org/ticket/5489
> > > > > > >
> > > > > > > It''s 1 year old already, but still
valid imho.
> > > > > >
> > > > > > I don''t quite follow what the problem is
with the current
> > > > > > implementation?  could you give me an example of
what bugs /
> > > > > > mis-behaviour this patch will prevent?
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Cheers
> > > > > >
> > > > > > Koz
> > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > > > >
> > > > >
> > > >
> > > >
> > > > --
> > > > Cheers
> > > >
> > > > Koz
> > > >
> > > > > > > >
> > > >
> > >
> >
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Core" group.
To post to this group, send email to rubyonrails-core@googlegroups.com
To unsubscribe from this group, send email to
rubyonrails-core-unsubscribe@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/rubyonrails-core?hl=en
-~----------~----~----~----~------~----~------~--~---