Resolves #106: Implementing skip-empty-xacts.

[bahaa]

Adding new parameter skip-empty-xacts. I chose the same parameter name as Postgres' test_decoding.

This should resolve #106.

[joekohlsdorf]

What is the reasoning behind not defaulting this to true?

[bahaa]

@joekohlsdorf backward compatibility. I think the plugin should behave the same for current users. Current users already get empty transactions and I think they are dealing with them already.

[joekohlsdorf]

In this particular case I would argue that the users aren't doing anything with empty transactions. So there is nothing lost by removing them because they contain no information at all.
They are just wasting network bandwidth and CPU cycles, depending on the use case this can require a substantial amount of additional processing.

[qweeze]

Any chance of merging this PR?

[eulerto]

@qweeze this PR is incomplete. It should be implemented in both format. @bahaa do you want to update this PR? Also, use a single commit.

[bahaa]

@eulerto sure. I can update it on the weekend. What do you mean by "both formats"?

[jmealo]

@bahaa: there's an option format-version which can be either 1 or 2 (see README) -- anything new needs to support both versions.

[bahaa]

@jmealo Thank you. I forked my version before there was support for two formats. I'll try to fix it this weekend.

[ajitj]

Looking forward to getting this feature. @bahaa any update on this?
Thanks

[llehtinen]

@bahaa would be great to get this merged

[jehanzebqayyum]

@bahaa pls merge this.

[adrijshikhar]

@bahaa if you are unable to work on this, I will be more than happy to take it further.

[bahaa]

@adrijshikhar sure. I'd be so glad if you do.

[adrijshikhar]

sure i will pick this up

[adrijshikhar]

#248