Re: [UPDATE] devel/ipython 5.3.0p0 -> 5.8.0 and devel/py-ipykernel 4.6.1p0 -> 4.9.0
OK by me to do the update, there is at least a new version of
py-ipykernel, but if is not a problem I would prefer to push this
commit, fix the problem with the dependencies of ipykernel and ipython
and then update each for separate.
Cheers.
Elias.
El lun., 29 oct. 2018 a las 1:18, Björn Ketelaars
() escribió:
>
> On Thu 06/09/2018 13:17, Elias M. Mariani wrote:
> > You are right, I forgot to attach the diff...
> > Here is the change in py-ipykernel that you requested.
> >
> > Cheers.
> > Elias.
> >
> > 2018-08-30 14:39 GMT-03:00 Elias M. Mariani :
> > > 2018-08-30 6:09 GMT-03:00 Edd Barrett :
> > >> Hi Elias,
> > >>
> > >> I've installed and ran ipython for Python 2 and 3. Seems to be working.
> > >> I've not tested any QT stuff -- I'll take your word that that works.
> > >>
> > >> On Wed, Aug 29, 2018 at 11:37:35PM -0300, Elias M. Mariani wrote:
> > >>> The self-test returns 1 error, I think that is related to
> > >>> graphics/py-matplotlib.
> > >>> (I added matplotlib to the test depends, it wasn't there before)
> > >>
> > >> And does that make the test pass?
> > > Sorry, maybe I didn't explain this in the right way:
> > > The test was running without grapchis/matplotlib, I added the
> > > dependency to enable the full extent of the tests, before this the
> > > test was skipped, so, this reveals a problem that most likely was
> > > hidden by the lack of the test dependency.
> > > The error was something like: "cannot write mode RGBA as JPEG" or
> > > something of the sort, given by python-PIL, while running matplotlib
> > > invoked by IPython, it doesn't look like a problem from IPython.
> > >
> > >>
> > >>> I disabled the building of the docs, I will send in a later mail
> > >>> devel/ipython-docs with the same content
> > >>
> > >> That's annoying, but I don't object in the interim. It would be nice to
> > >> have the docs back soon after this goes in though.
> > > I will commit that port as soon as this gets committed.
> > >
> > >>
> > >>> Might generate some problem with the Qt version used by some consumer.
> > >>
> > >> Do we have any QT consumers in our tree? I didn't think so.
> > >>
> > >> By the way:
> > >> ---8<---
> > >> +RUN_DEPENDS += devel/ipython${MODPY_FLAVOR}>=4.0.0 \
> > >> + devel/py-jupyter_client${MODPY_FLAVOR} \
> > >> + devel/py-traitlets${MODPY_FLAVOR}>=4.1.0 \
> > >> + www/py-tornado${MODPY_FLAVOR}>=4.0
> > >>
> > >> -RUN_DEPENDS += devel/py-jupyter_client${MODPY_FLAVOR}>=5.0.1 \
> > >> - devel/py-traitlets${MODPY_FLAVOR} \
> > >> - www/py-tornado${MODPY_FLAVOR}
> > >> --->8---
> > >>
> > >> If you move the devel/ipython line you added to the end, then it will be
> > >> easier to see what changed in the diff.
> > > Note taken.
> > >
> > >>
> > >> --
> > >> Best Regards
> > >> Edd Barrett
> > >>
> > >> http://www.theunixzoo.co.uk
>
>
> OK bket@ for both py-ipykernel-4.9.0, and py-ipython-5.3.0.
Re: [UPDATE] devel/ipython 5.3.0p0 -> 5.8.0 and devel/py-ipykernel 4.6.1p0 -> 4.9.0
On Thu 06/09/2018 13:17, Elias M. Mariani wrote:
> You are right, I forgot to attach the diff...
> Here is the change in py-ipykernel that you requested.
>
> Cheers.
> Elias.
>
> 2018-08-30 14:39 GMT-03:00 Elias M. Mariani :
> > 2018-08-30 6:09 GMT-03:00 Edd Barrett :
> >> Hi Elias,
> >>
> >> I've installed and ran ipython for Python 2 and 3. Seems to be working.
> >> I've not tested any QT stuff -- I'll take your word that that works.
> >>
> >> On Wed, Aug 29, 2018 at 11:37:35PM -0300, Elias M. Mariani wrote:
> >>> The self-test returns 1 error, I think that is related to
> >>> graphics/py-matplotlib.
> >>> (I added matplotlib to the test depends, it wasn't there before)
> >>
> >> And does that make the test pass?
> > Sorry, maybe I didn't explain this in the right way:
> > The test was running without grapchis/matplotlib, I added the
> > dependency to enable the full extent of the tests, before this the
> > test was skipped, so, this reveals a problem that most likely was
> > hidden by the lack of the test dependency.
> > The error was something like: "cannot write mode RGBA as JPEG" or
> > something of the sort, given by python-PIL, while running matplotlib
> > invoked by IPython, it doesn't look like a problem from IPython.
> >
> >>
> >>> I disabled the building of the docs, I will send in a later mail
> >>> devel/ipython-docs with the same content
> >>
> >> That's annoying, but I don't object in the interim. It would be nice to
> >> have the docs back soon after this goes in though.
> > I will commit that port as soon as this gets committed.
> >
> >>
> >>> Might generate some problem with the Qt version used by some consumer.
> >>
> >> Do we have any QT consumers in our tree? I didn't think so.
> >>
> >> By the way:
> >> ---8<---
> >> +RUN_DEPENDS += devel/ipython${MODPY_FLAVOR}>=4.0.0 \
> >> + devel/py-jupyter_client${MODPY_FLAVOR} \
> >> + devel/py-traitlets${MODPY_FLAVOR}>=4.1.0 \
> >> + www/py-tornado${MODPY_FLAVOR}>=4.0
> >>
> >> -RUN_DEPENDS += devel/py-jupyter_client${MODPY_FLAVOR}>=5.0.1 \
> >> - devel/py-traitlets${MODPY_FLAVOR} \
> >> - www/py-tornado${MODPY_FLAVOR}
> >> --->8---
> >>
> >> If you move the devel/ipython line you added to the end, then it will be
> >> easier to see what changed in the diff.
> > Note taken.
> >
> >>
> >> --
> >> Best Regards
> >> Edd Barrett
> >>
> >> http://www.theunixzoo.co.uk
OK bket@ for both py-ipykernel-4.9.0, and py-ipython-5.3.0.
Re: [UPDATE] devel/ipython 5.3.0p0 -> 5.8.0 and devel/py-ipykernel 4.6.1p0 -> 4.9.0
Sorry for ping this but I really would like to update devel/spyder. 2018-09-07 13:27 GMT-03:00 Elias M. Mariani : > OK edd@ > > 2018-09-07 12:39 GMT-03:00 Edd Barrett : >> On Thu, Sep 06, 2018 at 01:17:02PM -0300, Elias M. Mariani wrote: >>> You are right, I forgot to attach the diff... >>> Here is the change in py-ipykernel that you requested. >> >> LGTM. OK edd@ >> >> -- >> Best Regards >> Edd Barrett >> >> http://www.theunixzoo.co.uk
Re: [UPDATE] devel/ipython 5.3.0p0 -> 5.8.0 and devel/py-ipykernel 4.6.1p0 -> 4.9.0
OK edd@ 2018-09-07 12:39 GMT-03:00 Edd Barrett : > On Thu, Sep 06, 2018 at 01:17:02PM -0300, Elias M. Mariani wrote: >> You are right, I forgot to attach the diff... >> Here is the change in py-ipykernel that you requested. > > LGTM. OK edd@ > > -- > Best Regards > Edd Barrett > > http://www.theunixzoo.co.uk
Re: [UPDATE] devel/ipython 5.3.0p0 -> 5.8.0 and devel/py-ipykernel 4.6.1p0 -> 4.9.0
You are right, I forgot to attach the diff...
Here is the change in py-ipykernel that you requested.
Cheers.
Elias.
2018-08-30 14:39 GMT-03:00 Elias M. Mariani :
> 2018-08-30 6:09 GMT-03:00 Edd Barrett :
>> Hi Elias,
>>
>> I've installed and ran ipython for Python 2 and 3. Seems to be working.
>> I've not tested any QT stuff -- I'll take your word that that works.
>>
>> On Wed, Aug 29, 2018 at 11:37:35PM -0300, Elias M. Mariani wrote:
>>> The self-test returns 1 error, I think that is related to
>>> graphics/py-matplotlib.
>>> (I added matplotlib to the test depends, it wasn't there before)
>>
>> And does that make the test pass?
> Sorry, maybe I didn't explain this in the right way:
> The test was running without grapchis/matplotlib, I added the
> dependency to enable the full extent of the tests, before this the
> test was skipped, so, this reveals a problem that most likely was
> hidden by the lack of the test dependency.
> The error was something like: "cannot write mode RGBA as JPEG" or
> something of the sort, given by python-PIL, while running matplotlib
> invoked by IPython, it doesn't look like a problem from IPython.
>
>>
>>> I disabled the building of the docs, I will send in a later mail
>>> devel/ipython-docs with the same content
>>
>> That's annoying, but I don't object in the interim. It would be nice to
>> have the docs back soon after this goes in though.
> I will commit that port as soon as this gets committed.
>
>>
>>> Might generate some problem with the Qt version used by some consumer.
>>
>> Do we have any QT consumers in our tree? I didn't think so.
>>
>> By the way:
>> ---8<---
>> +RUN_DEPENDS += devel/ipython${MODPY_FLAVOR}>=4.0.0 \
>> + devel/py-jupyter_client${MODPY_FLAVOR} \
>> + devel/py-traitlets${MODPY_FLAVOR}>=4.1.0 \
>> + www/py-tornado${MODPY_FLAVOR}>=4.0
>>
>> -RUN_DEPENDS += devel/py-jupyter_client${MODPY_FLAVOR}>=5.0.1 \
>> - devel/py-traitlets${MODPY_FLAVOR} \
>> - www/py-tornado${MODPY_FLAVOR}
>> --->8---
>>
>> If you move the devel/ipython line you added to the end, then it will be
>> easier to see what changed in the diff.
> Note taken.
>
>>
>> --
>> Best Regards
>> Edd Barrett
>>
>> http://www.theunixzoo.co.uk
py-ipykernel.diff
Description: Binary data
Re: [UPDATE] devel/ipython 5.3.0p0 -> 5.8.0 and devel/py-ipykernel 4.6.1p0 -> 4.9.0
2018-08-30 6:09 GMT-03:00 Edd Barrett :
> Hi Elias,
>
> I've installed and ran ipython for Python 2 and 3. Seems to be working.
> I've not tested any QT stuff -- I'll take your word that that works.
>
> On Wed, Aug 29, 2018 at 11:37:35PM -0300, Elias M. Mariani wrote:
>> The self-test returns 1 error, I think that is related to
>> graphics/py-matplotlib.
>> (I added matplotlib to the test depends, it wasn't there before)
>
> And does that make the test pass?
Sorry, maybe I didn't explain this in the right way:
The test was running without grapchis/matplotlib, I added the
dependency to enable the full extent of the tests, before this the
test was skipped, so, this reveals a problem that most likely was
hidden by the lack of the test dependency.
The error was something like: "cannot write mode RGBA as JPEG" or
something of the sort, given by python-PIL, while running matplotlib
invoked by IPython, it doesn't look like a problem from IPython.
>
>> I disabled the building of the docs, I will send in a later mail
>> devel/ipython-docs with the same content
>
> That's annoying, but I don't object in the interim. It would be nice to
> have the docs back soon after this goes in though.
I will commit that port as soon as this gets committed.
>
>> Might generate some problem with the Qt version used by some consumer.
>
> Do we have any QT consumers in our tree? I didn't think so.
>
> By the way:
> ---8<---
> +RUN_DEPENDS += devel/ipython${MODPY_FLAVOR}>=4.0.0 \
> + devel/py-jupyter_client${MODPY_FLAVOR} \
> + devel/py-traitlets${MODPY_FLAVOR}>=4.1.0 \
> + www/py-tornado${MODPY_FLAVOR}>=4.0
>
> -RUN_DEPENDS += devel/py-jupyter_client${MODPY_FLAVOR}>=5.0.1 \
> - devel/py-traitlets${MODPY_FLAVOR} \
> - www/py-tornado${MODPY_FLAVOR}
> --->8---
>
> If you move the devel/ipython line you added to the end, then it will be
> easier to see what changed in the diff.
Note taken.
>
> --
> Best Regards
> Edd Barrett
>
> http://www.theunixzoo.co.uk
Re: [UPDATE] devel/ipython 5.3.0p0 -> 5.8.0 and devel/py-ipykernel 4.6.1p0 -> 4.9.0
Hi Elias,
I've installed and ran ipython for Python 2 and 3. Seems to be working.
I've not tested any QT stuff -- I'll take your word that that works.
On Wed, Aug 29, 2018 at 11:37:35PM -0300, Elias M. Mariani wrote:
> The self-test returns 1 error, I think that is related to
> graphics/py-matplotlib.
> (I added matplotlib to the test depends, it wasn't there before)
And does that make the test pass?
> I disabled the building of the docs, I will send in a later mail
> devel/ipython-docs with the same content
That's annoying, but I don't object in the interim. It would be nice to
have the docs back soon after this goes in though.
> Might generate some problem with the Qt version used by some consumer.
Do we have any QT consumers in our tree? I didn't think so.
By the way:
---8<---
+RUN_DEPENDS += devel/ipython${MODPY_FLAVOR}>=4.0.0 \
+ devel/py-jupyter_client${MODPY_FLAVOR} \
+ devel/py-traitlets${MODPY_FLAVOR}>=4.1.0 \
+ www/py-tornado${MODPY_FLAVOR}>=4.0
-RUN_DEPENDS += devel/py-jupyter_client${MODPY_FLAVOR}>=5.0.1 \
- devel/py-traitlets${MODPY_FLAVOR} \
- www/py-tornado${MODPY_FLAVOR}
--->8---
If you move the devel/ipython line you added to the end, then it will be
easier to see what changed in the diff.
--
Best Regards
Edd Barrett
http://www.theunixzoo.co.uk
[UPDATE] devel/ipython 5.3.0p0 -> 5.8.0 and devel/py-ipykernel 4.6.1p0 -> 4.9.0
%gui qt now uses Qt 5 by default rather than Qt 4, following a similar change in terminal IPython.devel/ipython and devel/py-ipykernel must get patched simultaneously to avoid generating a circular dependency. - devel/ipython: Changelog https://ipython.readthedocs.io/en/stable/whatsnew/version5.html#ipython-5-8-0 I see no possible problems, self-tested and tested some cosumers: devel/py-ipykernel (updated) shells/py-qtconsole www/jupyter-notebook The self-test returns 1 error, I think that is related to graphics/py-matplotlib. (I added matplotlib to the test depends, it wasn't there before) I disabled the building of the docs, I will send in a later mail devel/ipython-docs with the same content, but given that the docs need ipykernel (that depends on ipython to run) to build makes hard (meaning, not without lots of additional logic and work) to build ipython and ipython-docs altogether. We should wait on edd@ to decide if this is OK, given that he is the maintainer. - devel/py-ipykernel: Changelog https://github.com/ipython/ipykernel/blob/master/docs/changelog.rst I see no possible problems, except for: "%gui qt now uses Qt 5 by default rather than Qt 4, following a similar change in terminal IPython." Might generate some problem with the Qt version used by some consumer. Regression tests passing. Tested OK in consumers: shells/py-qtconsole www/jupyter-notebook Taking maintainership. (already talked with Alexandr) Cheers. Elias. ipython.diff Description: Binary data py-ipykernel.diff Description: Binary data
