librelist archives

« back to archive

Introduction & question

Introduction & question

From:
Dipl.-Ing. Luca Ingianni
Date:
2011-09-24 @ 08:54
[To be honest, this post is less structured than I'd like. I decided to
throw it at you all anyway because I felt that my question is valid,
even though my confusion at this time is such that I'm not even really
sure how to pose it well.
My apologies for being confusing.]

Hi all,

I just subscribed, and I'd like to quickly introduce myself. My name is
Luca, I work out of Munich, Germany. I'm an aeronautical engineer by
trade, but as it happens, my work involves a lot of coding :)

Since we often work in a distributed fashion and need to review code, I
got interested in hg-review. I like that I can use it offline, and it
feels nicely clean and simple. Overall, my first impression of the
software is quite favorable :)
I'm still trying to find out whether hg-review is the right tool for us.
It seems like it would do the trick, but I need to try it some more to
make up my mind.

However, there's one crucial question I have:
I just read the documentation, and to be honest, I am still confused how
to actually use hg-review. What's the workflow you envision/use?

So, say Luca and Adrian both work on a project.
They have a central repo on an Internet server that they use to
synchronize their code. They communicate frequently, face-to-face, over
the phone and via email.

The review repo would obviously also live on that server, and they would
frequently pull it. Then what?
Of course, they both can review, comment, and sign off to their heart's
content, but I feel that there's no guidance. I can't really see what I
reviewed already. I can't see what's new. I can't see what I should
review (because it's urgent, or because Adrian is stumped and needs my
input, or whatever). I don't see what comments Adrian added since my
last visit. I don't really get a quick overview over what the state of
the repo is, or what my next steps (as a developer, as a reviewer, as a
repo manager, ...) should be.

I'm not even sure yet what exactly it is that I'm missing. 
I suppose I don't have understood yet how to use hg-review in practice,
and the software itself lacks the kind of guidance that, for instance, a
bugtracker gives me: a sense of open issues and plausible next steps.
Quite possibly, I don't need guidance by the software, but just by the
documentation.

To reiterate: this is not intended as a "complaint" towards hg-review.
It's just what my immediate feeling is upon trying it out.

So, how do you all use it?
Are the functions I just hinted at even necessary, or am I making it too
complicated for myself?

Re: [hg.review] Introduction & question

From:
Haszlakiewicz, Eric
Date:
2011-09-28 @ 18:36
> -----Original Message-----
> From: hg.review@librelist.com [mailto:hg.review@librelist.com] On
> Behalf Of Dipl.-Ing. Luca Ingianni
> 
> So, say Luca and Adrian both work on a project.
> They have a central repo on an Internet server that they use to
> synchronize their code. They communicate frequently, face-to-face, over
> the phone and via email.
> 
> The review repo would obviously also live on that server, and they
> would
> frequently pull it. Then what?

Yep, the way I set it up was with the url for any review repositories 
being almost the same as the regular repository, but with a /hg_review/ 
instead of /hg/ in the path.

> Of course, they both can review, comment, and sign off to their heart's
> content, but I feel that there's no guidance. I can't really see what I
> reviewed already. I can't see what's new. I can't see what I should
> review (because it's urgent, or because Adrian is stumped and needs my
> input, or whatever). I don't see what comments Adrian added since my
> last visit. I don't really get a quick overview over what the state of
> the repo is, or what my next steps (as a developer, as a reviewer, as a
> repo manager, ...) should be.

Yeah, a lot of this is missing.  I've been using hg-review for a little 
while now and I've noticed the same things.
The way I've been treating it is that any changeset that doesn't have a 
signoff is one that needs to be reviewed.
Any changeset that has a "No" signoff is one that the change author needs 
to look at.

The workflow of how you notice that there's something for you to do still 
needs work.  I've just been sending emails manually after performing a 
review of a change, and periodically read through the comments for recent 
changes to see if there's anything that got missed.  Obviously, this is 
less than ideal.  Care to come up with a design for how to fix it?

I was thinking of making some improvements, starting with a few things 
that are considerably simpler than what you mentioned above, but I haven't
seen any response from Steve Losh (the maintainer) about actually getting 
changes into the main sources, so I've been rather reluctant to do much.  
:(

eric

Re: [hg.review] Introduction & question

From:
Luca Ingianni
Date:
2011-09-30 @ 13:10
Eric,

thanks a lot for your mail.

I wrote a huge reply yesterday, but my webmailer swallowed it - session
timed out :'(
I'll rewrite my reply soon (but not today, I'm still recovering from
Oktoberfest).

Luca