Website build process should delete existing HTML files

[orbitbot]

Since the website is indexed by Google and the like, end users may end up navigating to stale data, see chat @ https://gitter.im/lhorie/mithril.js?at=58c6bed600c00c3d4ffee6b0 and discussions before that.

I have the nagging feeling this might already have been implemented, if so sorry about the extra issue...

[dead-claudia]

I've already pointed this out to @lhorie, but pushing a fresh copy each time would probably be best.

[tivac]

The trick with this is that old version docs in /archive need to stick around, but everything else should be blown up. I'll see what I can figure out.

[tivac]

There isn't a good way to do this w/ the gh-pages package we're using to copy things to that branch. Would need to switch to handling it manually, bleh.

[dead-claudia]

@tivac I've always found it easier to just write a custom hack to generate simple sites like these.

[tivac]

@isiahmeadows I am averse to writing code if I can avoid it. If this continues to be a problem I can do the work to make it happen, but gonna table this for now.

[dead-claudia]

@tivac Metalsmith + a custom clean might work.

[tivac]

I don't want to recreate the entire process, I'll need to replace the gh-pages module with some actual git commands to properly take the generated ./dist folder and splat it onto the gh-pages branch.

[dead-claudia]

@tivac What about the archive directory?

Although, to be fair, I'm just throwing ideas at the wall. I don't have much investment other than it working. 😉

[tivac]

Yeah, that's the annoying bit that requires I not use the built-in cleaning functionality in the gh-pages module.

😠

[dead-claudia]

@tivac Just filed tschaub/gh-pages#180.

[StephanHoyer]

@orbitbot is this still an issue? Can you check plz?