ZIP destination consistency

[adam-lynch]

For a second, forget what should be in inside the ZIPs.

If you have a ZIP called a.zip, with just b.js in it, what's the path of b.js after unpacking? What's desired?

tempDirectory/b.js? tempDirectory/a/b.js?

Since e281927 and e41c864, on Windows, the contents of the ZIP is unpacked into a directory named after the basename of the ZIP (inside temp). So it would be tempDirectory/a/b.js.

If the other platforms don't do this, we should make it so.

[edjafarov]

yes I agree that tempDirectory/a/b.js makes sense for windows and linux.
the mac have app as a folder appName.app which is ok as it is.

[adam-lynch]

the mac have app as a folder appName.app which is ok as it is.

Does that apply to just when you're using DMGs or ZIps as well?

[edjafarov]

zip as well. The app in mac is a folder. You can run it. open appName.app though it is also a folder and you can just change it's content.

[adam-lynch]

Oh, but I mean they could still have a ZIP with a/b/c.app inside. I've making these sort of changes on my custom-names branch.

[edjafarov]

they may have. Though to copy stuff properly we have a restriction that there should be a single folder in the package at a root level. For win and linux we have several files generated by webkit-builder thus they have to be in a folder. For mac we have single folder as executable thus it may be at a root level.

[adam-lynch]

Though to copy stuff properly we have a restriction that there should be a single folder in the package at a root level.

Well, we will always because it'll be inside a package (ZIP, etc.) and the ZIP's basename is used to create a folder.

Have a good look at #37 and tell if something doesn't make sense.

[edjafarov]

I guess with #37 it finally does makes sense, thanks!

[adam-lynch]

😄

[adam-lynch]

Closing since #37 is merged