Genshi PY3 status

[Nicolas Évrard]

Hello,

If I remember correctly, the only think blocking python 3 support in
genshi (through 2to3) was the release of 0.6. I might be wrong but I
think the release has never been done.

Are there bugs that need some work to accelerate the movement ?

[Simon Cross]

Post by Nicolas Évrard
If I remember correctly, the only think blocking python 3 support in
genshi (through 2to3) was the release of 0.6. I might be wrong but I
think the release has never been done.

Yes, that's relatively accurate except that the release with Python 3
support will be 0.7 (0.6 was released some time ago).

Post by Nicolas Évrard
Are there bugs that need some work to accelerate the movement ?

There is one critical outstanding bug that would need to be fixed
before a 0.7 release:

http://genshi.edgewall.org/query?status=assigned&status=new&status=reopened&order=priority&priority=critical&col=id&col=summary&col=priority&col=status&col=owner&col=type&col=milestone

There are also a few test breakages introduced by the latest Python
2.7 (a few Genshi tests needlessly rely on some corner case behaviour
of the HTMLParser -- Genshi itself isn't affect by this).

[Nicklas Börjesson]

Hi,
I have perhaps misused the bug reporting system, but this seems also to be
a problem at py3.2: http://genshi.edgewall.org/ticket/428
It seems to be a difference in the parsing.

//Nicklas

Post by Simon Cross

Post by Nicolas Évrard
If I remember correctly, the only think blocking python 3 support in
genshi (through 2to3) was the release of 0.6. I might be wrong but I
think the release has never been done.

Yes, that's relatively accurate except that the release with Python 3
support will be 0.7 (0.6 was released some time ago).

Post by Nicolas Évrard
Are there bugs that need some work to accelerate the movement ?

There is one critical outstanding bug that would need to be fixed
http://genshi.edgewall.org/query?status=assigned&status=new&status=reopened&order=priority&priority=critical&col=id&col=summary&col=priority&col=status&col=owner&col=type&col=milestone
There are also a few test breakages introduced by the latest Python
2.7 (a few Genshi tests needlessly rely on some corner case behaviour
of the HTMLParser -- Genshi itself isn't affect by this).

[Nicolas Évrard]

Post by Simon Cross

Post by Nicolas Évrard
If I remember correctly, the only think blocking python 3 support in
genshi (through 2to3) was the release of 0.6. I might be wrong but I
think the release has never been done.

Yes, that's relatively accurate except that the release with Python 3
support will be 0.7 (0.6 was released some time ago).

Post by Nicolas Évrard
Are there bugs that need some work to accelerate the movement ?

There is one critical outstanding bug that would need to be fixed
http://genshi.edgewall.org/query?status=assigned&status=new&status=reopened&order=priority&priority=critical&col=id&col=summary&col=priority&col=status&col=owner&col=type&col=milestone

I added a patch on the ticket with the suggestion of MikeP. All tests
passes.

Post by Simon Cross
There are also a few test breakages introduced by the latest Python
2.7 (a few Genshi tests needlessly rely on some corner case behaviour
of the HTMLParser -- Genshi itself isn't affect by this).

I made a ticket and a patch for those tests:

http://genshi.edgewall.org/ticket/529

[Nicolas Évrard]

Post by Nicolas Évrard

Post by Simon Cross

Post by Nicolas Évrard
If I remember correctly, the only think blocking python 3 support in
genshi (through 2to3) was the release of 0.6. I might be wrong but I
think the release has never been done.

Yes, that's relatively accurate except that the release with Python 3
support will be 0.7 (0.6 was released some time ago).

Post by Nicolas Évrard
Are there bugs that need some work to accelerate the movement ?

There is one critical outstanding bug that would need to be fixed
http://genshi.edgewall.org/query?status=assigned&status=new&status=reopened&order=priority&priority=critical&col=id&col=summary&col=priority&col=status&col=owner&col=type&col=milestone

I added a patch on the ticket with the suggestion of MikeP. All tests
passes.

Post by Simon Cross
There are also a few test breakages introduced by the latest Python
2.7 (a few Genshi tests needlessly rely on some corner case behaviour
of the HTMLParser -- Genshi itself isn't affect by this).

BTW applying both patches and running the tests works under python3!
--
Nicolas Évrard

B2CK SPRL
rue de Rotterdam, 4
4000 LiÚge
Belgium
Tel: [phone]
E-mail/Jabber: ***@b2ck.com
Website: http://www.b2ck.com/

[jerry]

Great, does this mean an official Genshi release supporting Python3 is
around the corner?

Cheers,
Jerry

Post by Nicolas Évrard

Post by Nicolas Évrard

Post by Simon Cross

Post by Nicolas Évrard
If I remember correctly, the only think blocking python 3 support in
genshi (through 2to3) was the release of 0.6. I might be wrong but I
think the release has never been done.

Yes, that's relatively accurate except that the release with Python 3
support will be 0.7 (0.6 was released some time ago).

Post by Nicolas Évrard
Are there bugs that need some work to accelerate the movement ?

There is one critical outstanding bug that would need to be fixed
http://genshi.edgewall.org/query?status=assigned&status=new&status=re...

I added a patch on the ticket with the suggestion of MikeP. All tests
passes.

Post by Simon Cross
There are also a few test breakages introduced by the latest Python
2.7 (a few Genshi tests needlessly rely on some corner case behaviour
of the HTMLParser -- Genshi itself isn't affect by this).

BTW applying both patches and running the tests works under python3!
--
Nicolas Évrard
B2CK SPRL
rue de Rotterdam, 4
4000 Liège
Belgium
Tel: [phone]
Website:http://www.b2ck.com/
 signature.asc
< 1KViewDownload

[Nicolas Évrard]

Post by jerry
Great, does this mean an official Genshi release supporting Python3 is
around the corner?

I certainly hope so, genshi being the last library blocking me from
publishing my patches for the support of python3 on Tryton.

I must say that for a short moment I investigated the idea of using
another templating engine. The work of porting templates from one
engine to another made me change my mind.

[Nicolas Évrard]

Post by Nicolas Évrard

Post by Nicolas Évrard

Post by Simon Cross

Post by Nicolas Évrard
If I remember correctly, the only think blocking python 3 support in
genshi (through 2to3) was the release of 0.6. I might be wrong but I
think the release has never been done.

Yes, that's relatively accurate except that the release with Python 3
support will be 0.7 (0.6 was released some time ago).

Post by Nicolas Évrard
Are there bugs that need some work to accelerate the movement ?

There is one critical outstanding bug that would need to be fixed
http://genshi.edgewall.org/query?status=assigned&status=new&status=reopened&order=priority&priority=critical&col=id&col=summary&col=priority&col=status&col=owner&col=type&col=milestone

I added a patch on the ticket with the suggestion of MikeP. All tests
passes.

Post by Simon Cross
There are also a few test breakages introduced by the latest Python
2.7 (a few Genshi tests needlessly rely on some corner case behaviour
of the HTMLParser -- Genshi itself isn't affect by this).

BTW applying both patches and running the tests works under python3!

Hello,

It has been almost two weeks since I submitted those patches and I had
no information about them.

Are there any other stuff needed to be done before the release?

[Nicolas Évrard]

Post by Nicolas Évrard

Post by Nicolas Évrard

Post by Nicolas Évrard

Post by Simon Cross
There is one critical outstanding bug that would need to be fixed
http://genshi.edgewall.org/query?status=assigned&status=new&status=reopened&order=priority&priority=critical&col=id&col=summary&col=priority&col=status&col=owner&col=type&col=milestone

I added a patch on the ticket with the suggestion of MikeP. All tests
passes.

Post by Simon Cross
There are also a few test breakages introduced by the latest Python
2.7 (a few Genshi tests needlessly rely on some corner case behaviour
of the HTMLParser -- Genshi itself isn't affect by this).

BTW applying both patches and running the tests works under python3!

Hello,
It has been almost two weeks since I submitted those patches and I had
no information about them.
Are there any other stuff needed to be done before the release?

No answer in one week.

Since I really would like to see python3 support for genshi I would be
interested if there is a way to make it happen.

Any pointer would be really appreciated.

[Alessandro Molina]

Hi Nicolas,
I'm quite interested too as we are working on Py3 support in
TurboGears2 and we have an already working branch using the currently
not released version of Genshi.

If there is away the TurboGears team can help let us know, we would be glad to.

Bests,
Alessandro

Post by Nicolas Évrard

Post by Nicolas Évrard

Post by Nicolas Évrard

Post by Nicolas Évrard

Post by Simon Cross
There is one critical outstanding bug that would need to be fixed
http://genshi.edgewall.org/query?status=assigned&status=new&status=reopened&order=priority&priority=critical&col=id&col=summary&col=priority&col=status&col=owner&col=type&col=milestone

I added a patch on the ticket with the suggestion of MikeP. All tests
passes.

Post by Simon Cross
There are also a few test breakages introduced by the latest Python
2.7 (a few Genshi tests needlessly rely on some corner case behaviour
of the HTMLParser -- Genshi itself isn't affect by this).

BTW applying both patches and running the tests works under python3!

Hello,
It has been almost two weeks since I submitted those patches and I had
no information about them.
Are there any other stuff needed to be done before the release?

No answer in one week.
Since I really would like to see python3 support for genshi I would be
interested if there is a way to make it happen.
Any pointer would be really appreciated.
--
(°> Nicolas Évrard
( ) Liège
`¯

[Nicolas Évrard]

Post by Alessandro Molina
Hi Nicolas,

Hello Alessandro,

Post by Alessandro Molina
I'm quite interested too as we are working on Py3 support in
TurboGears2 and we have an already working branch using the currently
not released version of Genshi.
If there is away the TurboGears team can help let us know, we would be glad to.

What do you think of working on a genshi3 module that we could
maintain somewhere on github/bitbucket and then release on pypi.

Our project would depend of genshi when using python2 and on genshi3
when using python3 to run the setup.py script.

We would sync the repositories with upstream which should not be that
difficult given that there was no commit since at least six months.

[David Fraser]

Post by Nicolas Évrard

Post by Alessandro Molina
Hi Nicolas,

Hello Alessandro,

Post by Alessandro Molina
I'm quite interested too as we are working on Py3 support in
TurboGears2 and we have an already working branch using the
currently
not released version of Genshi.
If there is away the TurboGears team can help let us know, we would be glad to.

What do you think of working on a genshi3 module that we could
maintain somewhere on github/bitbucket and then release on pypi.
Our project would depend of genshi when using python2 and on genshi3
when using python3 to run the setup.py script.
We would sync the repositories with upstream which should not be that
difficult given that there was no commit since at least six months.

Having genshi in a distributed vcs, even if subversion remains as the main source for now, would help a lot.
So far there's https://bitbucket.org/edgewall/genshi (Mercurial) and https://github.com/edgewall/genshi (git, but empty), and I've got https://github.com/davidfraser/genshi as a git-svn mirror, but it's not showing all the tags.
Is there an official plan for edgewall projects in general?

Cheers
David

[Alessandro Molina]

I would really prefer to have only one official package, especially
because a working version of genshi on python3 is probably not far
away considering that I have been fairly able to run most of my
TurboGears2 applications on Python3 using the genshi trunk.

Post by Nicolas Évrard

Post by Alessandro Molina
Hi Nicolas,

Hello Alessandro,

Post by Alessandro Molina
I'm quite interested too as we are working on Py3 support in
TurboGears2 and we have an already working branch using the currently
not released version of Genshi.
If there is away the TurboGears team can help let us know, we would be glad to.

What do you think of working on a genshi3 module that we could
maintain somewhere on github/bitbucket and then release on pypi.
Our project would depend of genshi when using python2 and on genshi3
when using python3 to run the setup.py script.
We would sync the repositories with upstream which should not be that
difficult given that there was no commit since at least six months.
--
(°> Nicolas Évrard
( ) Liège
`¯

[Nicolas Évrard]

Post by Alessandro Molina
I would really prefer to have only one official package, especially
because a working version of genshi on python3 is probably not far
away considering that I have been fairly able to run most of my
TurboGears2 applications on Python3 using the genshi trunk.

Well I agree with you but the lack of reactivity of the team is to say
the least worrying.

So either we choose to switch to another template engine (but with
some inconveniences, at least, for my users and probably also some
missing features) or we find a solution to make genshi usable under
python3.

After a few month, I decided to choose the last option, this solution
is to create another package but I am willing to accept any other
possibility.

[David Fraser]

Post by Nicolas Évrard

Post by Alessandro Molina
I would really prefer to have only one official package, especially
because a working version of genshi on python3 is probably not far
away considering that I have been fairly able to run most of my
TurboGears2 applications on Python3 using the genshi trunk.

Well I agree with you but the lack of reactivity of the team is to say
the least worrying.
So either we choose to switch to another template engine (but with
some inconveniences, at least, for my users and probably also some
missing features) or we find a solution to make genshi usable under
python3.
After a few month, I decided to choose the last option, this solution
is to create another package but I am willing to accept any other
possibility.

Are your issues only the actual release, or do you have other problems running Python3 on genshi trunk?

Cheers
David

[Nicolas Évrard]

Post by David Fraser

Post by Nicolas Évrard

Post by Alessandro Molina
I would really prefer to have only one official package, especially
because a working version of genshi on python3 is probably not far
away considering that I have been fairly able to run most of my
TurboGears2 applications on Python3 using the genshi trunk.

Well I agree with you but the lack of reactivity of the team is to
say the least worrying.
So either we choose to switch to another template engine (but with
some inconveniences, at least, for my users and probably also some
missing features) or we find a solution to make genshi usable under
python3.
After a few month, I decided to choose the last option, this solution
is to create another package but I am willing to accept any other
possibility.

Are your issues only the actual release, or do you have other
problems running Python3 on genshi trunk?

So far I don't have any problem running python3 on genshi trunk.

But making release is good for software : it makes them alive. pypi is
also a big part of the python infrastructure and without a genshi
release we're blocking every framework/library from switching to
python3.

[Richard Mitchell]

Post by Nicolas Évrard
So far I don't have any problem running python3 on genshi trunk.
But making release is good for software : it makes them alive. pypi is
also a big part of the python infrastructure and without a genshi
release we're blocking every framework/library from switching to
python3.

Same here, no problems with Python 3.2 on trunk.

What's the hold up? Better to have a Python 3 release out there and people
using / testing against it than no release at all if you're waiting for it
to be perfect. This is blocking many packages downstream from supporting
Python 3.