add --prepend option as altenative to prepend-token

[mansona]

Hi, me again 👋

Thanks again for answering my issue #165 so quickly, and with implementation no less 😱

I noticed you implemented this idea of a "prepend token" which would make it quite challenging to automate the way that I'm hoping to 🤔 it means that after we make a release with a prepend token, we would need to move the prepend token to the top of the file again. Also if anyone forgets to do this while I'm not managing the project then they run the risk of deleting other people's changelogs.

I added a --prepend option that effectively acts like there is a prepend-token at the start of the file at all times. Let me know if you have any questions or if you need me to do anything else to get this merged

[cookpete]

I'm curious as to how this works, assuming your changelog has a header or title? Would this not just be adding more and more headers every time you generate?

I feel like instead of --prepend you're better off using --stdout, then you can just pipe the output wherever you like?

[mansona]

actually that is a good point... I can see now why you implemented the prepend-token, I didn't understand it before.

Every changelog that I have does not have a dedicated header or paragraph at the top of the changelog. Here is an example of the one I've been testing auto-changelog on: https://github.com/ember-learn/ember-styleguide/blob/master/CHANGELOG.md

Is it common to have titles or descriptions on changelogs? Most of the time I see changelogs as just a list like the one I linked 🤔

Yea I guess you're right about --stdout although trying to prepend to a file with piping can be a bit of a pain and is potentially error prone

[WoodLeg]

Could be awesome to add this feature 👍

[cookpete]

Released in 2.2.0.