Proper default value for skipLimit
Recently I stumbled over the default value of the skipLimit in the class FaultTolerantStepBuilder. It is set by default to 0. If you don't use a SkipPolicy and configure it instead with skip(Exception.class) it is very easy to forget setting the additional property skipLimit. Could you provide a more suitable default value? My suggestion would be Integer.MAX_VALUE. What do you think?
That's a valid point, thank you for raising it!
My suggestion would be Integer.MAX_VALUE
This means that by default we would theoretically allow the entire dataset to be skipped, which in turn means there is something fundamentally wrong with the input (bad data format, unexpected input, etc). As a user, I would prefer my job to fail fast in the first dozen or hundred skips rather than waiting until the end just to discover that my job was skipping the entire data set. Do you see my point?
I guess 10 or 100 is a reasonable default. Wdyt?
I will plan the change in the next minor release.
Hi @fmbenhassine, this makes abolutely sense to me. I would prefer at least 10. Then you have a good balance between fail-fast and misconfiguration.
I have opened a pull request that addresses this issue: #4668