Lennart Regebro kirjoitti:

2009/9/23 Alex Grönholm :

...

Yes, but did you or did you not commit such changes to the repository?

Yes, I did. Martin v Löwis did most of the work, so if you are looking for my username you won't find them.

http://bitbucket.org/tarek/distribute/src/9145ea8b647b/

I'm talking about the changes that would add the three keyword arguments to setup() (use_2to3 etc) so we wouldn't have to deal with two different APIs. Such changes have not yet been committed to Tarek's repo. In my opinion, 0.6.2 should not be released before this is done.

...

At least easy_install does not have the --user option added, which is necessary for this to work.

Huh? It's not specifically a part of the Python 3 support, so if it's not added I'm not sure why you think the Python 3 support should add it. Can you explain what exactly you are trying to do, what you expect to happen and what actually happens

I just wanted to install distribute (and other packages as well) as non-root so I wouldn't screw up the system in case something went badly wrong. You're right of course, about it not being an inherent part of Python 3 support.

Lennart Regebro kirjoitti:

2009/9/23 Alex Grönholm :

...

I'm talking about the changes that would add the three keyword arguments to setup() (use_2to3 etc) so we wouldn't have to deal with two different APIs. Such changes have not yet been committed to Tarek's repo. In my opinion, 0.6.2 should not be released before this is done.

They have been commited, since a week, by Martin v Löwis. The push I did yesterday to rename them apparently didn't work. hg is still confusing me.

I pushed again.

Yup, I see it now. It seems you accidentally created a new head two days ago. Shame on me for not noticing that earlier (bitbucket visualizes those things pretty well in the changesets view).

...

I just wanted to install distribute (and other packages as well) as non-root so I wouldn't screw up the system in case something went badly wrong.

Quite understandable, but it seems that's not in 0.6.x, it doesn't show up under Linux either. Somebody else have to answer about this.

I know it's not in 0.6.x -- the question is, should it be added in 0.6.2, or in a future 0.6.x release, or left completely to 0.7?

Lennart Regebro kirjoitti:

2009/9/23 Alex Grönholm :

...

Yup, I see it now. It seems you accidentally created a new head two days ago.

I did? How did I do that? How are you supposed to cimmit/push without doing that, and why doesn't it bloody fail?

It doesn't fail because it's not an error. I'm not sure if I'm interpreting the graph right, but every time I'm about to accidentally do that, Eclipse screams bloody murder, so I don't think that was me.

In short: Why does hg suck so much? I am extremely close to actually refuse to use it until it's usable without a doctors degree in version management.

What would you rather use? Subversion? Let's face it, distributed SCMs aren't the easiest thing on the planet to learn, but they are the best solution for open source projects.

Lennart Regebro kirjoitti:

2009/9/23 Alex Grönholm :

...

Yes, but did you or did you not commit such changes to the repository?

Yes, I did. Martin v Löwis did most of the work, so if you are looking for my username you won't find them.

http://bitbucket.org/tarek/distribute/src/9145ea8b647b/

...

At least easy_install does not have the --user option added, which is necessary for this to work.

Huh? It's not specifically a part of the Python 3 support, so if it's not added I'm not sure why you think the Python 3 support should add it. Can you explain what exactly you are trying to do, what you expect to happen and what actually happens?

Just to clarify, I had not specified setuptools.run_2to3 = True anywhere, which is why any packages I installed on py3k were installed as-is, so no problems there (yet).