On 23 May 2011 10:10, Brett Porter <[email protected]> wrote: > Wow, serendipity. I was going to implement that later tonight :) >
Yeah well the Cassandra guys need retries in gpg:sign-and-deploy-file so symmetry demands the same in deploy:deploy-file > Nice one! > > Just a couple of comments inline: > > On 23/05/2011, at 6:02 PM, [email protected] wrote: > >> >> [MDEPLOY-133] > > I think the log comment could cover more :) >> >> + catch ( ArtifactDeploymentException e ) >> + { > > Are there specific types of exceptions that are "fatal" that can be captured? > 502/503 responses make sense to retry, but not sure about 500, 403, 401, etc. > you assume that all deployments are over Http(s)... and there may be a case whereby I realise that I don't have permissions in the middle of the release and I run to fix them, just missing the first attempt ;-) >> + if (count + 1 < retryFailedDeploymentCount) { >> + getLog().warn( "Something went wrong with the >> deployment, will try again", e ); > > This may be more verbose than necessary - how about just e.getMessage, with > the trace logged at debug level? > Done: r1126407/8 >> + } >> + if ( exception == null ) >> + { >> + exception = e; >> + } > > A completely new exception that says to refer to the other failures might be > more appropriate, if they happen to differ... though that may be too much of > an edge case :) > well if you want to push that shared exception up to a common module between m-gpg-p and m-deploy-p fine by me ;-) > Cheers, > Brett > > -- > Brett Porter > [email protected] > http://brettporter.wordpress.com/ > http://au.linkedin.com/in/brettporter > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
