Quantic phenomena in migrations

Rails
1

I have an application with 15 migrations under version control. In a
Mac and and in a Windows, a rake migrate from scratch runs them all
just fine. But in a different Windows machine rake migrate stops
after migration 3 for no apparent reason. --trace seems normal. No
error is reported. Both Windows are XP SP2. They all have the same
svn revision and Rails-related software, database is SQLite.

I even rebooted the machine to no avail. Does this ring any bell?

– fxn

2

Xavier N. <fxn@…> writes:

I have an application with 15 migrations under version control. In a
Mac and and in a Windows, a rake migrate from scratch runs them all
just fine. But in a different Windows machine rake migrate stops
after migration 3 for no apparent reason. --trace seems normal. No
error is reported. Both Windows are XP SP2. They all have the same
svn revision and Rails-related software, database is SQLite.

I even rebooted the machine to no avail. Does this ring any bell?

No bells, but I would re-checkout on Windows machine 2 into a clean
directory,
clear your database and try again. Perhaps something is screwed up with
the
state of those version control files.

-damon

3

On Jul 17, 2006, at 15:34, Damon C. wrote:

No bells, but I would re-checkout on Windows machine 2 into a clean
directory,
clear your database and try again. Perhaps something is screwed up
with the
state of those version control files.

Yeah, I tried without luck. I even checked the disk just in case
because this morning an Ubuntu in a different partition run a dosfsck
at startup. Nothing, lost by now.

4

On Jul 17, 2006, at 16:41, Xavier N. wrote:

Yeah, I tried without luck. I even checked the disk just in case
because this morning an Ubuntu in a different partition run a
dosfsck at startup. Nothing, lost by now.

Ah, let me add something I discovered. The problem seems to be
related to migration 4, the one that does not get to be run. If I
remove that file everything runs, until the 15th, the last one. I
can’t for the life of me see anything wrong in that migration, which
otherwise runs smoothly in two machines. I copy it below just for the
record.

– fxn

class LaCaixaCrawlerStatus < ActiveRecord::Migration
def self.up
delete_all
CrawlerStatus.create(:entity => Entity::LA_CAIXA)
end

def self.down
delete_all
end

def self.delete_all
CrawlerStatus.delete_all([‘entity = ?’, Entity::LA_CAIXA])
end
end

5

On Jul 17, 2006, at 10:47 AM, Xavier N. wrote:

which otherwise runs smoothly in two machines. I copy it below just
def self.down
delete_all
end

def self.delete_all
CrawlerStatus.delete_all([‘entity = ?’, Entity::LA_CAIXA])
end
end

How has your CrawlerStatus model changed between when migrations 4
and 15 were created? This is a “known” problem when using Model
classes in migrations. (particularly when your model deals with more
columns/attributes than the table has at that point in time)

-Rob

Rob B. http://agileconsultingllc.com
[email protected]

6

Just for the archives.

After a whole day lost doing all kind of trials, changing SQLite with
MySQL, removing rails from vendor, doing clean checkouts, and what
not, I decided to reinstall Windows, this is a quite new laptop so
that was an option. Now the migrations run normally again. The only
strange thing I did this morning was to boot a brand new Kubuntu that
automatically runs dosfsck (that does not mean it was its fault, but
is the only conjecture I have).

Unfortunately I don’t know what happened, but at least I’ll be able
to continue with this project tomorrow.

– fxn