> In some cases we use other utilities to inference changes to files 
> besides hg, for example, OpenGrok and webrev.  In the case for Opengrok, 
> it truncates all but the first line of comments in the changeset, so 
> that the comments are of no relevance of the associated file at all.  

That would seem to be a bug against OpenGrok, which should be using the 
description in place of the summary for each changeset.

> In the case for webrev, it lumps all the comments for every file thereby 
> confusing the reviewer of which is the relevant CR.

Webrev doesn't lump anything together.  It reports what is in the 
repository under review.  If the webrev is generated from a repository 
that has been recommitted, you'll see what you describe.  But if it's 
generated (as seems more appropriate) from a repository with a valid 
change history, it will report comments sensibly.

> A solution, as I think Milan already suggested, could include multiple 
> changesets in a workspace, but alas this is cumbersome to manage (viz 
> re-commits).  But I'm sure someone will create a clever tool to easily 
> manage multiple changesets in a ws, it's just a matter of time :)

The cadmium tools don't support this.

There are ways to achieve it, now.  Use of export/import works 
gracefully, as does using the mq extension.

The pain of learning Mercurial cannot, as I have previously stated, be 
avoided.  We're not going back to "the old model."

But we're also not shoving additional tools down the throats of the 
development community.

--Mark


Reply via email to