3 ms·

I apologize for that - at the time I misunderstood Mercurial internals. It confused me that Hg does store all it's files this way - and importantly it stores i

by schacon 17y ago

I apologize for that - at the time I misunderstood Mercurial internals. It confused me that Hg does store all it's files this way - and importantly it stores its file revisions under the file name, so renames have to be tracked explicitly to find the history of the file properly, which isn't the case with Git. I was focused on that - content vs file tracking. However, you are right, Mercurial is fundamentally a DAG based system and I didn't know that at the time and nobody I had talked to about this knew either. To be fair, just a few months ago a Googler made the inverse mistake about Git when talking about Hg at Google IO, so I'm not alone. :)

gecko 17y ago
Quite all right. :) I think a lot of people locked into Git or Mercurial have a distorted view of how different the two systems are. Especially these days, the differences are ultimately minor, which is why projects like hg-git, which allows transparently working with a Git repositories from Mercurial, are possible in the first place. Yes, Mercurial's revlogs versus Git's blobs have performance implications for both systems, but that difference is not meaningfully user-visible 99% of the time--and the rest really just ends up being differences in terminology and commands, more than anything.