On Jun 19, 2008, at 1:44 PM, David Jencks wrote:


On Jun 19, 2008, at 10:26 AM, Kevan Miller wrote:


On Jun 19, 2008, at 12:27 PM, Lin Sun wrote:

I think the current exception is rather confusion when a user tries to
assemble a server from admin console.   He will see an error message
saying the file was not found in the server's console which makes him
to think the created custom assembly zip file is incorrect.

How about we change to - when the file is not found, log it as a
warning or info?   For other exceptions, log it as an error as
currently the code is doing.

Lin

On Thu, Jun 19, 2008 at 11:59 AM, David Jencks <[EMAIL PROTECTED] > wrote:
I agree with Jarek, -1 on this change. The warnings during the build are fairly harmless. Ideally we could figure out how to make the file be there before we try to access it during server assembly. I'd be be ok with logging an error but slightly prefer the current exception unless we find
there is no good way to create the file before it is needed.

They look like "errors" to me ;-P. And disagree that they are harmless. As described in the jira, this is not just during a "build" (i.e. while running mvn). It happens when exporting an assembly from a running server. I think users would justifiably think that something has gone wrong.

That said, I also agree that it's better to inform users when we don't find a config-substitutions.properties file.

The appropriate solution, IMO, is to include a config- substitutions.properties file in geronimo-boilerplate-minimal.jar. I think everyone will be happy, then.

Not me :-), at least not yet. Unfortunately I don't remember what eventually creates this file. I thought I once got rid of this error message by making sure the file was created at the appropriate time.

My original idea was that since a particular config- substitutions.properties type file was defined by a server instance gbean we shouldn't create any such files unless we knew how they should be named, from a server-instance gbean.

I'm in the middle of some other stuff and don't want to change gears today.... I would start investigating by figuring out what does eventually create the files and why that isn't run first during assembly.

Sounds good. As long as we have this specific problem resolved, I'll prolly be happy with the specifics of the fix.

--kevan

Reply via email to