I would agree that testing the migration should be a part of the
development process, but I think of it as a level of testing that comes
after my software. First and foremost, the software has to work with
the version of the database that it is designed to run against. I do
understand that wi
My concern with the changeset-level preconditions and updateSql mode is
that since the database isn't actually updated, most of the time the
preconditions will fail for no valid reason.
For example:
In update mode, changeSet B will run, but in updateSql mode, the
precondi
Just wanted to let you know that I have submitted a patch for bug 2688450
(updateSQL ignores precondition & generate SQL even though PreCondition
fails). I have also attached it to this message, since I could not attach a
file in tracker and was forced to paste the patch into a comment.
By al
> Developer A should have merged from the SCM, then done a complete
> re-test, including rebuilding the database. At least, that's how I see it.
Yep, that is another solution I'd forgotten, but it is also one I was trying to
avoid. In general, I would not recommend rebuilding the database bec
It looks to me like the problem in this scenario is NOT recreating a
database from the ground up for testing. I constantly struggle with
this problem with our QE department. They want to keep pulling their
data forward as changes occur, instead of creating a script that will
regenerate their
Hi all,
here is an interesting case that happened during our Liquibase/changeSet
development. I was already aware of this possibility, but this was the first
time that it really happened in our environment.
In short, the problem is that, when developers hold their changeSets locally
for a few
I don't care much about this, but I think that anyone who still needs Java 1.4
can download a previous version of Liquibase and use it, right?
Diego
- Mensagem original
> De: Paul Keeble
> Para: liquibase-user@lists.sourceforge.net
> Enviadas: Segunda-feira, 20 de Abril de 2009 11:1
If only it was that easy! 1.4 is the backbone of huge numbers of application
servers still in active use out there, because JavaEE only recently got
updated. These are also the environments where a tool like liquibase comes in
most useful. I'd love to see the death of Java 1.4 as much as every
Java 1.4 has been end-of-lifed: "The EOL transition period began Dec, 11 2006
and will complete
October 30th, 2008"
http://java.sun.com/j2se/1.4.2/
Let's not IE6 this -- kill it, kill it now, kill it dead, and no more support.
~~ Robert.
Christian Maslen wrote:
>
>
> Voxland, Nathan wrote: