Re: ANN: SciPy 1.0 beta release

classic Classic list List threaded Threaded
17 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Joe Harrington
On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <[hidden email]> wrote:

> I'm excited to be able to announce the availability of the first beta
> release of Scipy 1.0.
>
[I won't quote the entire release notes!]

Awesome news, looooong in the making!

The number of deprecations and removals is understandably large.
I think we need a centralized resource for navigating deprecations,
removals, and changes.  It could be a simple web page, or a more
involved database.  It should list the change, give or link to the
rationale, and state the date and package version number.  Most
importantly, it should give the suggested replacement(s).

Googling "scipy deprecation" found only this page, plus a bunch of
release notes and discussions.  The page was last updated on Jan 16,
2017, and is limited to C-API deprecations:

https://docs.scipy.org/doc/numpy-dev/reference/c-api.deprecations.html

Even once such a document/page/searchable database exists, many will be
unaware of it, or not know where to find it, so referring to it at the
top of any list of deprecations/removals would be good.

Once it exists, deprecations could be entered before being implemented,
and deprecation warnings could refer to an index number in this
resource, where the user could go for more information on how to fix the
issue.

Thanks,

--jh--
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

ralfgommers


On Mon, Sep 18, 2017 at 1:16 AM, Joe Harrington <[hidden email]> wrote:
On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <[hidden email]> wrote:

> I'm excited to be able to announce the availability of the first beta
> release of Scipy 1.0.
>
[I won't quote the entire release notes!]

Awesome news, looooong in the making!

The number of deprecations and removals is understandably large.
I think we need a centralized resource for navigating deprecations,
removals, and changes.  It could be a simple web page, or a more
involved database.  It should list the change, give or link to the
rationale, and state the date and package version number.  Most
importantly, it should give the suggested replacement(s).

Googling "scipy deprecation" found only this page, plus a bunch of
release notes and discussions.  The page was last updated on Jan 16,
2017, and is limited to C-API deprecations:

https://docs.scipy.org/doc/numpy-dev/reference/c-api.deprecations.html

Even once such a document/page/searchable database exists, many will be
unaware of it, or not know where to find it, so referring to it at the
top of any list of deprecations/removals would be good.

Once it exists, deprecations could be entered before being implemented,
and deprecation warnings could refer to an index number in this
resource, where the user could go for more information on how to fix the
issue.

Thanks for your thoughts Joe! I'm not sure a database is the way to go, but you make a good point. We have to guide users a bit more in what to do about deprecations. I think clear instructions about upgrading in both the html docs and release announcements would be useful. I don't think there's many users who run ``python -Wd my_code.py`` .(making depreacations that affect your code visible) ....

Ralf


Thanks,

--jh--
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user


_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

ralfgommers
In reply to this post by Joe Harrington


On Sun, Sep 17, 2017 at 10:48 PM, Ralf Gommers <[hidden email]> wrote:
Hi all,

I'm excited to be able to announce the availability of the first beta release of Scipy 1.0. This is a big release, and a version number that has been 16 years in the making. It contains a few more deprecations and backwards incompatible changes than an average release. Therefore please do test it on your own code, and report any issues on the Github issue tracker or on the scipy-dev mailing list.

Binary wheels: will follow tomorrow, I'll announce those when ready (TravisCI is under maintenance right now)

Binary wheels for Windows, Linux and OS X (for all supported Python versions, 32-bit and 64-bit) can be found at http://wheels.scipy.org. To install directly with pip:

    pip install scipy=='1.0.0b1' -f http://wheels.scipy.org --trusted-host wheels.scipy.org

(add --user and/or --upgrade as required to that command). Alternatively, just download the wheel you need and do `pip install scipy-1.0.0b1-<downloaded>.whl`.

Cheers,
Ralf



_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Thomas Kluyver-2
On 18 September 2017 at 10:59, Ralf Gommers <[hidden email]> wrote:
Binary wheels for Windows, Linux and OS X (for all supported Python versions, 32-bit and 64-bit) can be found at http://wheels.scipy.org. To install directly with pip:

    pip install scipy=='1.0.0b1' -f http://wheels.scipy.org --trusted-host wheels.scipy.org

I don't want to criticise the hard work that has gone into making this available, but I'm disappointed that we're telling people to install software over an insecure HTTP connection.

Can the wheels not be uploaded to PyPI?

_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Matthew Brett
Hi,

On Mon, Sep 18, 2017 at 11:07 AM, Thomas Kluyver <[hidden email]> wrote:

> On 18 September 2017 at 10:59, Ralf Gommers <[hidden email]> wrote:
>>
>> Binary wheels for Windows, Linux and OS X (for all supported Python
>> versions, 32-bit and 64-bit) can be found at http://wheels.scipy.org. To
>> install directly with pip:
>>
>>     pip install scipy=='1.0.0b1' -f http://wheels.scipy.org --trusted-host
>> wheels.scipy.org
>
>
> I don't want to criticise the hard work that has gone into making this
> available, but I'm disappointed that we're telling people to install
> software over an insecure HTTP connection.

I personally prefer the following recipe:

pip install -f https://3f23b170c54c2533c070-1c8a9b3114517dc5fe17b7c3f8c63a43.ssl.cf2.rackcdn.com
scipy=='1.0.0b1'

> Can the wheels not be uploaded to PyPI?

Sounds like a good idea.  I can do that - any objections?

Cheers,

Matthew
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

ralfgommers
In reply to this post by Thomas Kluyver-2


On Mon, Sep 18, 2017 at 10:07 PM, Thomas Kluyver <[hidden email]> wrote:
On 18 September 2017 at 10:59, Ralf Gommers <[hidden email]> wrote:
Binary wheels for Windows, Linux and OS X (for all supported Python versions, 32-bit and 64-bit) can be found at http://wheels.scipy.org. To install directly with pip:

    pip install scipy=='1.0.0b1' -f http://wheels.scipy.org --trusted-host wheels.scipy.org

I don't want to criticise the hard work that has gone into making this available, but I'm disappointed that we're telling people to install software over an insecure HTTP connection.

Can the wheels not be uploaded to PyPI?

I thought we'd never done that before, but I'm wrong: https://pypi.python.org/pypi/scipy/0.18.0rc2

So yes, I think.

Ralf




_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

ralfgommers
In reply to this post by Matthew Brett


On Mon, Sep 18, 2017 at 10:11 PM, Matthew Brett <[hidden email]> wrote:
Hi,

On Mon, Sep 18, 2017 at 11:07 AM, Thomas Kluyver <[hidden email]> wrote:
> On 18 September 2017 at 10:59, Ralf Gommers <[hidden email]> wrote:
>>
>> Binary wheels for Windows, Linux and OS X (for all supported Python
>> versions, 32-bit and 64-bit) can be found at http://wheels.scipy.org. To
>> install directly with pip:
>>
>>     pip install scipy=='1.0.0b1' -f http://wheels.scipy.org --trusted-host
>> wheels.scipy.org
>
>
> I don't want to criticise the hard work that has gone into making this
> available, but I'm disappointed that we're telling people to install
> software over an insecure HTTP connection.

I personally prefer the following recipe:

pip install -f https://3f23b170c54c2533c070-1c8a9b3114517dc5fe17b7c3f8c63a43.ssl.cf2.rackcdn.com
scipy=='1.0.0b1'

> Can the wheels not be uploaded to PyPI?

Sounds like a good idea.  I can do that - any objections?

That would be helpful Matthew, I'm about to sign off for today.

Cheers,
Ralf



_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Matthew Brett
Hi,

On Mon, Sep 18, 2017 at 11:14 AM, Ralf Gommers <[hidden email]> wrote:

>
>
> On Mon, Sep 18, 2017 at 10:11 PM, Matthew Brett <[hidden email]>
> wrote:
>>
>> Hi,
>>
>> On Mon, Sep 18, 2017 at 11:07 AM, Thomas Kluyver <[hidden email]> wrote:
>> > On 18 September 2017 at 10:59, Ralf Gommers <[hidden email]>
>> > wrote:
>> >>
>> >> Binary wheels for Windows, Linux and OS X (for all supported Python
>> >> versions, 32-bit and 64-bit) can be found at http://wheels.scipy.org.
>> >> To
>> >> install directly with pip:
>> >>
>> >>     pip install scipy=='1.0.0b1' -f http://wheels.scipy.org
>> >> --trusted-host
>> >> wheels.scipy.org
>> >
>> >
>> > I don't want to criticise the hard work that has gone into making this
>> > available, but I'm disappointed that we're telling people to install
>> > software over an insecure HTTP connection.
>>
>> I personally prefer the following recipe:
>>
>> pip install -f
>> https://3f23b170c54c2533c070-1c8a9b3114517dc5fe17b7c3f8c63a43.ssl.cf2.rackcdn.com
>> scipy=='1.0.0b1'
>>
>> > Can the wheels not be uploaded to PyPI?
>>
>> Sounds like a good idea.  I can do that - any objections?
>
>
> That would be helpful Matthew, I'm about to sign off for today.

Done - new instructions for testing:

pip install --pre --upgrade scipy

Cheers,

Matthew
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

ralfgommers
In reply to this post by ralfgommers


On Mon, Sep 18, 2017 at 8:58 PM, Ralf Gommers <[hidden email]> wrote:


On Mon, Sep 18, 2017 at 1:16 AM, Joe Harrington <[hidden email]> wrote:
On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <[hidden email]> wrote:

> I'm excited to be able to announce the availability of the first beta
> release of Scipy 1.0.
>
[I won't quote the entire release notes!]

Awesome news, looooong in the making!

The number of deprecations and removals is understandably large.
I think we need a centralized resource for navigating deprecations,
removals, and changes.  It could be a simple web page, or a more
involved database.  It should list the change, give or link to the
rationale, and state the date and package version number.  Most
importantly, it should give the suggested replacement(s).

Googling "scipy deprecation" found only this page, plus a bunch of
release notes and discussions.  The page was last updated on Jan 16,
2017, and is limited to C-API deprecations:

https://docs.scipy.org/doc/numpy-dev/reference/c-api.deprecations.html

Even once such a document/page/searchable database exists, many will be
unaware of it, or not know where to find it, so referring to it at the
top of any list of deprecations/removals would be good.

Once it exists, deprecations could be entered before being implemented,
and deprecation warnings could refer to an index number in this
resource, where the user could go for more information on how to fix the
issue.

Thanks for your thoughts Joe! I'm not sure a database is the way to go, but you make a good point. We have to guide users a bit more in what to do about deprecations. I think clear instructions about upgrading in both the html docs and release announcements would be useful. I don't think there's many users who run ``python -Wd my_code.py`` .(making depreacations that affect your code visible) ....


Ralf



_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Thomas Jollans
In reply to this post by ralfgommers
On 2017-09-18 10:58, Ralf Gommers wrote:
> I don't think there's many users who run ``python -Wd my_code.py``.
> (making depreacations that affect your code visible) ....


I'm sure they don't, especially seeing as it's not obvious from the
Python docs that "-Wd" does *anything*: requesting the default behaviour
*should* be a noop...

However, the default for IPython is to display DeprecationsWarning and
PendingDeprecationWarning, so quite a lot of people will actually see them.


--
Thomas Jollans
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Joe Harrington
In reply to this post by ralfgommers
Um...  This commit has to do with wheels, not deprecations.  Great that
we're flying with wheels up, though. :-)

--jh--

On Thu, 28 Sep 2017 21:48:41 +1300,  Ralf Gommers <[hidden email]>
wrote:

> Done in https://github.com/scipy/scipy/pull/7947

>> On Mon, Sep 18, 2017 at 1:16 AM, Joe Harrington <[hidden email]>
>> wrote:
>>
>>> On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <[hidden email]>
>>> wrote:
>>>
>>> > I'm excited to be able to announce the availability of the first beta
>>> > release of Scipy 1.0.
>>> >
>>> [I won't quote the entire release notes!]
>>>
>>> Awesome news, looooong in the making!
>>>
>>> The number of deprecations and removals is understandably large.
>>> I think we need a centralized resource for navigating deprecations,
>>> removals, and changes.  It could be a simple web page, or a more
>>> involved database.  It should list the change, give or link to the
>>> rationale, and state the date and package version number.  Most
>>> importantly, it should give the suggested replacement(s).
>>>
>>> Googling "scipy deprecation" found only this page, plus a bunch of
>>> release notes and discussions.  The page was last updated on Jan 16,
>>> 2017, and is limited to C-API deprecations:
>>>
>>> https://docs.scipy.org/doc/numpy-dev/reference/c-api.deprecations.html
>>>
>>> Even once such a document/page/searchable database exists, many will be
>>> unaware of it, or not know where to find it, so referring to it at the
>>> top of any list of deprecations/removals would be good.
>>>
>>> Once it exists, deprecations could be entered before being implemented,
>>> and deprecation warnings could refer to an index number in this
>>> resource, where the user could go for more information on how to fix the
>>> issue.
>>>
>>
>> Thanks for your thoughts Joe! I'm not sure a database is the way to go,
>> but you make a good point. We have to guide users a bit more in what to do
>> about deprecations. I think clear instructions about upgrading in both the
>> html docs and release announcements would be useful. I don't think there's
>> many users who run ``python -Wd my_code.py`` .(making depreacations that
>> affect your code visible) ....
>>
>
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Joe Harrington
Ok, I saw the -Wd at the end, but it just requests the default, so how
does this change anything?  If it's just an ipython benefit, say that
it's recommended for ipython users.

Also, remember that for the vast majority of users, they don't install a
new version, they suddenly discover that their system now has one,
through some software update.  Think students and casual users, as
opposed to PhD research scientists experienced in software management.

But, what we really need is more information about what to DO about the
deprecations.  Not, "we pulled the rug out from under you", but "so that
we don't pull the rug out from under you, do this".  And, in a
more-obvious place than the release notes, which most users don't ever
see.

--jh--

>Um...  This commit has to do with wheels, not deprecations.  Great that
>we're flying with wheels up, though. :-)
>
>--jh--
>
>On Thu, 28 Sep 2017 21:48:41 +1300,  Ralf Gommers <[hidden email]>
>wrote:
>
>> Done in https://github.com/scipy/scipy/pull/7947
>
>>> On Mon, Sep 18, 2017 at 1:16 AM, Joe Harrington <[hidden email]>
>>> wrote:
>>>
>>>> On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <[hidden email]>
>>>> wrote:
>>>>
>>>> > I'm excited to be able to announce the availability of the first beta
>>>> > release of Scipy 1.0.
>>>> >
>>>> [I won't quote the entire release notes!]
>>>>
>>>> Awesome news, looooong in the making!
>>>>
>>>> The number of deprecations and removals is understandably large.
>>>> I think we need a centralized resource for navigating deprecations,
>>>> removals, and changes.  It could be a simple web page, or a more
>>>> involved database.  It should list the change, give or link to the
>>>> rationale, and state the date and package version number.  Most
>>>> importantly, it should give the suggested replacement(s).
>>>>
>>>> Googling "scipy deprecation" found only this page, plus a bunch of
>>>> release notes and discussions.  The page was last updated on Jan 16,
>>>> 2017, and is limited to C-API deprecations:
>>>>
>>>> https://docs.scipy.org/doc/numpy-dev/reference/c-api.deprecations.html
>>>>
>>>> Even once such a document/page/searchable database exists, many will be
>>>> unaware of it, or not know where to find it, so referring to it at the
>>>> top of any list of deprecations/removals would be good.
>>>>
>>>> Once it exists, deprecations could be entered before being implemented,
>>>> and deprecation warnings could refer to an index number in this
>>>> resource, where the user could go for more information on how to fix the
>>>> issue.
>>>>
>>>
>>> Thanks for your thoughts Joe! I'm not sure a database is the way to go,
>>> but you make a good point. We have to guide users a bit more in what to do
>>> about deprecations. I think clear instructions about upgrading in both the
>>> html docs and release announcements would be useful. I don't think there's
>>> many users who run ``python -Wd my_code.py`` .(making depreacations that
>>> affect your code visible) ....
>>>
>>
>
>
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Warren Weckesser-2


On Thu, Sep 28, 2017 at 9:44 AM, Joe Harrington <[hidden email]> wrote:
Ok, I saw the -Wd at the end, but it just requests the default, so how
does this change anything?


Here's a fun sentence from the documentation at https://docs.python.org/3.6/library/warnings.html#updating-code-for-new-versions-of-python regarding the option -Wd:

> This enables default handling for all warnings, including those that are ignored by default.

Fortunately, the sentences previous to that one explain that when -Wd is used, DeprecationWarnings (usually ignored) are made visible.

Warren


If it's just an ipython benefit, say that
it's recommended for ipython users.

Also, remember that for the vast majority of users, they don't install a
new version, they suddenly discover that their system now has one,
through some software update.  Think students and casual users, as
opposed to PhD research scientists experienced in software management.

But, what we really need is more information about what to DO about the
deprecations.  Not, "we pulled the rug out from under you", but "so that
we don't pull the rug out from under you, do this".  And, in a
more-obvious place than the release notes, which most users don't ever
see.

--jh--

>Um...  This commit has to do with wheels, not deprecations.  Great that
>we're flying with wheels up, though. :-)
>
>--jh--
>
>On Thu, 28 Sep 2017 21:48:41 +1300,  Ralf Gommers <[hidden email]>
>wrote:
>
>> Done in https://github.com/scipy/scipy/pull/7947
>
>>> On Mon, Sep 18, 2017 at 1:16 AM, Joe Harrington <[hidden email]>
>>> wrote:
>>>
>>>> On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <[hidden email]>
>>>> wrote:
>>>>
>>>> > I'm excited to be able to announce the availability of the first beta
>>>> > release of Scipy 1.0.
>>>> >
>>>> [I won't quote the entire release notes!]
>>>>
>>>> Awesome news, looooong in the making!
>>>>
>>>> The number of deprecations and removals is understandably large.
>>>> I think we need a centralized resource for navigating deprecations,
>>>> removals, and changes.  It could be a simple web page, or a more
>>>> involved database.  It should list the change, give or link to the
>>>> rationale, and state the date and package version number.  Most
>>>> importantly, it should give the suggested replacement(s).
>>>>
>>>> Googling "scipy deprecation" found only this page, plus a bunch of
>>>> release notes and discussions.  The page was last updated on Jan 16,
>>>> 2017, and is limited to C-API deprecations:
>>>>
>>>> https://docs.scipy.org/doc/numpy-dev/reference/c-api.deprecations.html
>>>>
>>>> Even once such a document/page/searchable database exists, many will be
>>>> unaware of it, or not know where to find it, so referring to it at the
>>>> top of any list of deprecations/removals would be good.
>>>>
>>>> Once it exists, deprecations could be entered before being implemented,
>>>> and deprecation warnings could refer to an index number in this
>>>> resource, where the user could go for more information on how to fix the
>>>> issue.
>>>>
>>>
>>> Thanks for your thoughts Joe! I'm not sure a database is the way to go,
>>> but you make a good point. We have to guide users a bit more in what to do
>>> about deprecations. I think clear instructions about upgrading in both the
>>> html docs and release announcements would be useful. I don't think there's
>>> many users who run ``python -Wd my_code.py`` .(making depreacations that
>>> affect your code visible) ....
>>>
>>
>
>
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user


_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Joe Harrington
All of this is orthogonal to the original and much more important point:
We're not telling people HOW to fix their deprecated code.  Perhaps an
interested student could volunteer to make a page on scipy.org listing
the deprecations and removals in each version (pulled from the release
notes).  Then, that person, developers, or other interested users could
add modern alternatives and the rationale for removal.

To the minor point about -Wd, is scipy turning off DeprecationWarnings
by default, or is some environment like ipython doing it?

I see numpy deprecation warnings all over the place, until I fix them.
They're not turned off by default.  I'm running in straight Python on
Ubuntu Linux, with numpy and sometimes scipy.  I'm not using ipython.
This behavior has been true in all versions of Python, numpy, and scipy
I've used.

So, I think that paragraph could use another look, from the perspective
of not just the preferred environment of the developer who wrote it.  If
SciPy turns off DeprecationWarnings and numpy doesn't, does this really
make sense?  Because numpy doesn't.

If it's an environment doing it, then it should be noted that -Wd is
needed only in those environments.  Finally, it should be noted that -Wd
is shorthand for -W default, which is all the Python man page talks
about.  Searching for -Wd there comes up with nothing.

--jh--

On Thu, 28 Sep 2017 11:21:50 -0400, Warren Weckesser
<[hidden email]> wrote:

> On Thu, Sep 28, 2017 at 9:44 AM, Joe Harrington <[hidden email]> wrote:
>
> > Ok, I saw the -Wd at the end, but it just requests the default, so how
> > does this change anything?
>
>
>
> Here's a fun sentence from the documentation at
> https://docs.python.org/3.6/library/warnings.html#updating-code-for-new-versio>
> ns-of-python
> regarding the option -Wd:
>
> > This enables default handling for all warnings, including those that are
> ignored by default.
>
> Fortunately, the sentences previous to that one explain that when -Wd is
> used, DeprecationWarnings (usually ignored) are made visible.
>
> Warren
>
>
> If it's just an ipython benefit, say that
> > it's recommended for ipython users.
> >
> > Also, remember that for the vast majority of users, they don't install a
> > new version, they suddenly discover that their system now has one,
> > through some software update.  Think students and casual users, as
> > opposed to PhD research scientists experienced in software management.
> >
> > But, what we really need is more information about what to DO about the
> > deprecations.  Not, "we pulled the rug out from under you", but "so that
> > we don't pull the rug out from under you, do this".  And, in a
> > more-obvious place than the release notes, which most users don't ever
> > see.
> >
> > --jh--
> >
> > >Um...  This commit has to do with wheels, not deprecations.  Great that
> > >we're flying with wheels up, though. :-)
> > >
> > >--jh--
> > >
> > >On Thu, 28 Sep 2017 21:48:41 +1300,  Ralf Gommers <[hidden email]
> > >
> > >wrote:
> > >
> > >> Done in https://github.com/scipy/scipy/pull/7947
> > >
> > >>> On Mon, Sep 18, 2017 at 1:16 AM, Joe Harrington <[hidden email]>
> > >>> wrote:
> > >>>
> > >>>> On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <
> > [hidden email]>
> > >>>> wrote:
> > >>>>
> > >>>> > I'm excited to be able to announce the availability of the first
> > beta
> > >>>> > release of Scipy 1.0.
> > >>>> >
> > >>>> [I won't quote the entire release notes!]
> > >>>>
> > >>>> Awesome news, looooong in the making!
> > >>>>
> > >>>> The number of deprecations and removals is understandably large.
> > >>>> I think we need a centralized resource for navigating deprecations,
> > >>>> removals, and changes.  It could be a simple web page, or a more
> > >>>> involved database.  It should list the change, give or link to the
> > >>>> rationale, and state the date and package version number.  Most
> > >>>> importantly, it should give the suggested replacement(s).
> > >>>>
> > >>>> Googling "scipy deprecation" found only this page, plus a bunch of
> > >>>> release notes and discussions.  The page was last updated on Jan 16,
> > >>>> 2017, and is limited to C-API deprecations:
> > >>>>
> > >>>> https://docs.scipy.org/doc/numpy-dev/reference/c-api.
> > deprecations.html
> > >>>>
> > >>>> Even once such a document/page/searchable database exists, many will
> > be
> > >>>> unaware of it, or not know where to find it, so referring to it at the
> > >>>> top of any list of deprecations/removals would be good.
> > >>>>
> > >>>> Once it exists, deprecations could be entered before being
> > implemented,
> > >>>> and deprecation warnings could refer to an index number in this
> > >>>> resource, where the user could go for more information on how to fix
> > the
> > >>>> issue.
> > >>>>
> > >>>
> > >>> Thanks for your thoughts Joe! I'm not sure a database is the way to go,
> > >>> but you make a good point. We have to guide users a bit more in what
> > to do
> > >>> about deprecations. I think clear instructions about upgrading in both
> > the
> > >>> html docs and release announcements would be useful. I don't think
> > there's
> > >>> many users who run ``python -Wd my_code.py`` .(making depreacations
> > that
> > >>> affect your code visible) ....
> > >>>
> > >>
> > >
> > >
> > _______________________________________________
> > SciPy-User mailing list
> > [hidden email]
> > https://mail.python.org/mailman/listinfo/scipy-user
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Robert Kern-2
On Thu, Sep 28, 2017 at 10:56 AM, Joe Harrington <[hidden email]> wrote:

> To the minor point about -Wd, is scipy turning off DeprecationWarnings
> by default, or is some environment like ipython doing it?

It's a plain-old-python thing.

❯ python
Enthought Deployment Manager -- https://www.enthought.com
Python 2.7.13 |Enthought, Inc. (x86_64)| (default, Mar  2 2017, 08:20:50)
[GCC 4.2.1 Compatible Apple LLVM 6.0 (clang-600.0.57)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import warnings
>>> warnings.filters
[('ignore', None, <type 'exceptions.DeprecationWarning'>, None, 0), ('ignore', None, <type 'exceptions.PendingDeprecationWarning'>, None, 0), ('ignore', None, <type 'exceptions.ImportWarning'>, None, 0), ('ignore', None, <type 'exceptions.BytesWarning'>, None, 0)]

> I see numpy deprecation warnings all over the place, until I fix them.

You might be thinking of the FutureWarnings, which are not ignored by default.

--
Robert Kern

_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

ralfgommers
In reply to this post by Joe Harrington


On Fri, Sep 29, 2017 at 6:56 AM, Joe Harrington <[hidden email]> wrote:
All of this is orthogonal to the original and much more important point:
We're not telling people HOW to fix their deprecated code. 

Eh, we do. The "deprecated features" section of https://github.com/scipy/scipy/blob/master/doc/release/1.0.0-notes.rst has details for most items. In case where that becomes impractical like for all the misc.pilutil functions, the deprecation warnings themselves will say what to do. Example: https://github.com/scipy/scipy/blob/master/scipy/misc/pilutil.py#L108

The only exceptions are for things we remove because they're just wrong, like bytescale: https://github.com/scipy/scipy/blob/master/scipy/misc/pilutil.py#L32. There we can't tell the user what to use, because it depends on the application - but almost certainly bytescale is not the right solution. Hence it's just removed.
 
Perhaps an
interested student could volunteer to make a page on scipy.org listing
the deprecations and removals in each version (pulled from the release
notes).  Then, that person, developers, or other interested users could
add modern alternatives and the rationale for removal.

I'm -1 on a website for this on scipy.org, even if there was a volunteer. It's much less discoverable than having good warnings in the code, and will inevitably go out of date.

Ralf

 

To the minor point about -Wd, is scipy turning off DeprecationWarnings
by default, or is some environment like ipython doing it?

I see numpy deprecation warnings all over the place, until I fix them.
They're not turned off by default.  I'm running in straight Python on
Ubuntu Linux, with numpy and sometimes scipy.  I'm not using ipython.
This behavior has been true in all versions of Python, numpy, and scipy
I've used.

So, I think that paragraph could use another look, from the perspective
of not just the preferred environment of the developer who wrote it.  If
SciPy turns off DeprecationWarnings and numpy doesn't, does this really
make sense?  Because numpy doesn't.

If it's an environment doing it, then it should be noted that -Wd is
needed only in those environments.  Finally, it should be noted that -Wd
is shorthand for -W default, which is all the Python man page talks
about.  Searching for -Wd there comes up with nothing.

--jh--

On Thu, 28 Sep 2017 11:21:50 -0400, Warren Weckesser
<[hidden email]> wrote:

> On Thu, Sep 28, 2017 at 9:44 AM, Joe Harrington <[hidden email]> wrote:
>
> > Ok, I saw the -Wd at the end, but it just requests the default, so how
> > does this change anything?
>
>
>
> Here's a fun sentence from the documentation at
> https://docs.python.org/3.6/library/warnings.html#updating-code-for-new-versio>
> ns-of-python
> regarding the option -Wd:
>
> > This enables default handling for all warnings, including those that are
> ignored by default.
>
> Fortunately, the sentences previous to that one explain that when -Wd is
> used, DeprecationWarnings (usually ignored) are made visible.
>
> Warren
>
>
> If it's just an ipython benefit, say that
> > it's recommended for ipython users.
> >
> > Also, remember that for the vast majority of users, they don't install a
> > new version, they suddenly discover that their system now has one,
> > through some software update.  Think students and casual users, as
> > opposed to PhD research scientists experienced in software management.
> >
> > But, what we really need is more information about what to DO about the
> > deprecations.  Not, "we pulled the rug out from under you", but "so that
> > we don't pull the rug out from under you, do this".  And, in a
> > more-obvious place than the release notes, which most users don't ever
> > see.
> >
> > --jh--
> >
> > >Um...  This commit has to do with wheels, not deprecations.  Great that
> > >we're flying with wheels up, though. :-)
> > >
> > >--jh--
> > >
> > >On Thu, 28 Sep 2017 21:48:41 +1300,  Ralf Gommers <[hidden email]
> > >
> > >wrote:
> > >
> > >> Done in https://github.com/scipy/scipy/pull/7947
> > >
> > >>> On Mon, Sep 18, 2017 at 1:16 AM, Joe Harrington <[hidden email]>
> > >>> wrote:
> > >>>
> > >>>> On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <
> > [hidden email]>
> > >>>> wrote:
> > >>>>
> > >>>> > I'm excited to be able to announce the availability of the first
> > beta
> > >>>> > release of Scipy 1.0.
> > >>>> >
> > >>>> [I won't quote the entire release notes!]
> > >>>>
> > >>>> Awesome news, looooong in the making!
> > >>>>
> > >>>> The number of deprecations and removals is understandably large.
> > >>>> I think we need a centralized resource for navigating deprecations,
> > >>>> removals, and changes.  It could be a simple web page, or a more
> > >>>> involved database.  It should list the change, give or link to the
> > >>>> rationale, and state the date and package version number.  Most
> > >>>> importantly, it should give the suggested replacement(s).
> > >>>>
> > >>>> Googling "scipy deprecation" found only this page, plus a bunch of
> > >>>> release notes and discussions.  The page was last updated on Jan 16,
> > >>>> 2017, and is limited to C-API deprecations:
> > >>>>
> > >>>> https://docs.scipy.org/doc/numpy-dev/reference/c-api.
> > deprecations.html
> > >>>>
> > >>>> Even once such a document/page/searchable database exists, many will
> > be
> > >>>> unaware of it, or not know where to find it, so referring to it at the
> > >>>> top of any list of deprecations/removals would be good.
> > >>>>
> > >>>> Once it exists, deprecations could be entered before being
> > implemented,
> > >>>> and deprecation warnings could refer to an index number in this
> > >>>> resource, where the user could go for more information on how to fix
> > the
> > >>>> issue.
> > >>>>
> > >>>
> > >>> Thanks for your thoughts Joe! I'm not sure a database is the way to go,
> > >>> but you make a good point. We have to guide users a bit more in what
> > to do
> > >>> about deprecations. I think clear instructions about upgrading in both
> > the
> > >>> html docs and release announcements would be useful. I don't think
> > there's
> > >>> many users who run ``python -Wd my_code.py`` .(making depreacations
> > that
> > >>> affect your code visible) ....
> > >>>
> > >>
> > >
> > >
> > _______________________________________________
> > SciPy-User mailing list
> > [hidden email]
> > https://mail.python.org/mailman/listinfo/scipy-user
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user


_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user
Reply | Threaded
Open this post in threaded view
|

Re: ANN: SciPy 1.0 beta release

Thomas Jollans
In reply to this post by Joe Harrington
On 2017-09-28 19:56, Joe Harrington wrote:
> To the minor point about -Wd, is scipy turning off DeprecationWarnings
> by default, or is some environment like ipython doing it?

Python turns them off by default; ipython turns them ON (so a lot of
users actually see them)

> I see numpy deprecation warnings all over the place, until I fix them.
> They're not turned off by default.  I'm running in straight Python on
> Ubuntu Linux, with numpy and sometimes scipy.  I'm not using ipython.
> This behavior has been true in all versions of Python, numpy, and scipy
> I've used.

That's interesting, since the default Pythons on Ubuntu don't show
deprecation warnings:

0:tjol@krikkit:~% python
Python 2.7.13 (default, Jan 19 2017, 14:48:08)
[GCC 6.3.0 20170118] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import warnings
>>> warnings.warn(DeprecationWarning('foo'))
>>>
0:tjol@krikkit:~% python3
Python 3.5.3 (default, Jan 19 2017, 14:11:04)
[GCC 6.3.0 20170118] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import warnings
>>> warnings.warn(DeprecationWarning('foo'))
>>>
0:tjol@krikkit:~% ipython
Python 3.5.3 (default, Jan 19 2017, 14:11:04)
Type 'copyright', 'credits' or 'license' for more information
IPython 6.1.0 -- An enhanced Interactive Python. Type '?' for help.

In [1]: import warnings

In [2]: warnings.warn(DeprecationWarning('foo'))
/home/tjol/.local/bin/ipython:1: DeprecationWarning: foo
  #!/usr/bin/python3

In [3]:
Do you really want to exit ([y]/n)?
0:tjol@krikkit:~% python3 -Wd
Python 3.5.3 (default, Jan 19 2017, 14:11:04)
[GCC 6.3.0 20170118] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import warnings
>>> warnings.warn(DeprecationWarning('foo'))
__main__:1: DeprecationWarning: foo
>>>


>
> So, I think that paragraph could use another look, from the perspective
> of not just the preferred environment of the developer who wrote it.  If
> SciPy turns off DeprecationWarnings and numpy doesn't, does this really
> make sense?  Because numpy doesn't.
>
> If it's an environment doing it, then it should be noted that -Wd is
> needed only in those environments.  Finally, it should be noted that -Wd
> is shorthand for -W default, which is all the Python man page talks
> about.  Searching for -Wd there comes up with nothing.
>
> --jh--
>
> On Thu, 28 Sep 2017 11:21:50 -0400, Warren Weckesser
> <[hidden email]> wrote:
>
>> On Thu, Sep 28, 2017 at 9:44 AM, Joe Harrington <[hidden email]> wrote:
>>
>>> Ok, I saw the -Wd at the end, but it just requests the default, so how
>>> does this change anything?
>>
>>
>>
>> Here's a fun sentence from the documentation at
>> https://docs.python.org/3.6/library/warnings.html#updating-code-for-new-versio>
>> ns-of-python
>> regarding the option -Wd:
>>
>>> This enables default handling for all warnings, including those that are
>> ignored by default.
>>
>> Fortunately, the sentences previous to that one explain that when -Wd is
>> used, DeprecationWarnings (usually ignored) are made visible.
>>
>> Warren
>>
>>
>> If it's just an ipython benefit, say that
>>> it's recommended for ipython users.
>>>
>>> Also, remember that for the vast majority of users, they don't install a
>>> new version, they suddenly discover that their system now has one,
>>> through some software update.  Think students and casual users, as
>>> opposed to PhD research scientists experienced in software management.
>>>
>>> But, what we really need is more information about what to DO about the
>>> deprecations.  Not, "we pulled the rug out from under you", but "so that
>>> we don't pull the rug out from under you, do this".  And, in a
>>> more-obvious place than the release notes, which most users don't ever
>>> see.
>>>
>>> --jh--
>>>
>>>> Um...  This commit has to do with wheels, not deprecations.  Great that
>>>> we're flying with wheels up, though. :-)
>>>>
>>>> --jh--
>>>>
>>>> On Thu, 28 Sep 2017 21:48:41 +1300,  Ralf Gommers <[hidden email]
>>>>
>>>> wrote:
>>>>
>>>>> Done in https://github.com/scipy/scipy/pull/7947
>>>>
>>>>>> On Mon, Sep 18, 2017 at 1:16 AM, Joe Harrington <[hidden email]>
>>>>>> wrote:
>>>>>>
>>>>>>> On Sun, 17 Sep 2017 22:48:35 +1200 Ralf Gommers <
>>> [hidden email]>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> I'm excited to be able to announce the availability of the first
>>> beta
>>>>>>>> release of Scipy 1.0.
>>>>>>>>
>>>>>>> [I won't quote the entire release notes!]
>>>>>>>
>>>>>>> Awesome news, looooong in the making!
>>>>>>>
>>>>>>> The number of deprecations and removals is understandably large.
>>>>>>> I think we need a centralized resource for navigating deprecations,
>>>>>>> removals, and changes.  It could be a simple web page, or a more
>>>>>>> involved database.  It should list the change, give or link to the
>>>>>>> rationale, and state the date and package version number.  Most
>>>>>>> importantly, it should give the suggested replacement(s).
>>>>>>>
>>>>>>> Googling "scipy deprecation" found only this page, plus a bunch of
>>>>>>> release notes and discussions.  The page was last updated on Jan 16,
>>>>>>> 2017, and is limited to C-API deprecations:
>>>>>>>
>>>>>>> https://docs.scipy.org/doc/numpy-dev/reference/c-api.
>>> deprecations.html
>>>>>>>
>>>>>>> Even once such a document/page/searchable database exists, many will
>>> be
>>>>>>> unaware of it, or not know where to find it, so referring to it at the
>>>>>>> top of any list of deprecations/removals would be good.
>>>>>>>
>>>>>>> Once it exists, deprecations could be entered before being
>>> implemented,
>>>>>>> and deprecation warnings could refer to an index number in this
>>>>>>> resource, where the user could go for more information on how to fix
>>> the
>>>>>>> issue.
>>>>>>>
>>>>>>
>>>>>> Thanks for your thoughts Joe! I'm not sure a database is the way to go,
>>>>>> but you make a good point. We have to guide users a bit more in what
>>> to do
>>>>>> about deprecations. I think clear instructions about upgrading in both
>>> the
>>>>>> html docs and release announcements would be useful. I don't think
>>> there's
>>>>>> many users who run ``python -Wd my_code.py`` .(making depreacations
>>> that
>>>>>> affect your code visible) ....
>>>>>>
>>>>>
>>>>
>>>>
>>> _______________________________________________
>>> SciPy-User mailing list
>>> [hidden email]
>>> https://mail.python.org/mailman/listinfo/scipy-user
> _______________________________________________
> SciPy-User mailing list
> [hidden email]
> https://mail.python.org/mailman/listinfo/scipy-user
>
_______________________________________________
SciPy-User mailing list
[hidden email]
https://mail.python.org/mailman/listinfo/scipy-user