Menu
▾
▴
matplotlib-devel
|
From: Michael D. <md...@st...> - 2013-05-29 18:44:17
|
I'm pleased to announce the tagging of matplotlib-1.3.0rc1. Once the binaries from Christoph and Russell have been uploaded, I'll make a broader announcement to get some testing of this in advance of the final release. The tarball is available here: https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz The documentation for this version is viewable here: http://matplotlib.org/1.3.0 Thanks everyone for their hard work getting this out the door! Mike |
|
From: Christoph G. <cg...@uc...> - 2013-05-29 18:56:00
|
Hi, minor issue: the tarball contains a lib\matplotlib.egg-info directory. The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and therefore fails to build on my system. Deleting the lib\matplotlib.egg-info folder works. Does the release announcement mention that Windows binaries won't include required dependencies (pytz, dateutil, pyparsing, six) any longer? Christoph On 5/29/2013 11:41 AM, Michael Droettboom wrote: > I'm pleased to announce the tagging of matplotlib-1.3.0rc1. > > Once the binaries from Christoph and Russell have been uploaded, I'll > make a broader announcement to get some testing of this in advance of > the final release. > > The tarball is available here: > > https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz > > > The documentation for this version is viewable here: > > http://matplotlib.org/1.3.0 > > Thanks everyone for their hard work getting this out the door! > > Mike > |
|
From: Michael D. <md...@st...> - 2013-05-29 19:00:31
|
On 05/29/2013 02:55 PM, Christoph Gohlke wrote: > Hi, > > minor issue: the tarball contains a lib\matplotlib.egg-info directory. > The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and > therefore fails to build on my system. Deleting the > lib\matplotlib.egg-info folder works. Doh. This is the first time building with setuptools -- I'll fix that and create a new tarball... 1.3.0rc2 here we come! > > Does the release announcement mention that Windows binaries won't > include required dependencies (pytz, dateutil, pyparsing, six) any > longer? Yes. http://matplotlib.org/1.3.0/users/whats_new.html#new-setup-script > > Christoph > > > On 5/29/2013 11:41 AM, Michael Droettboom wrote: >> I'm pleased to announce the tagging of matplotlib-1.3.0rc1. >> >> Once the binaries from Christoph and Russell have been uploaded, I'll >> make a broader announcement to get some testing of this in advance of >> the final release. >> >> The tarball is available here: >> >> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz >> >> >> >> The documentation for this version is viewable here: >> >> http://matplotlib.org/1.3.0 >> >> Thanks everyone for their hard work getting this out the door! >> >> Mike >> |
|
From: Michael D. <md...@st...> - 2013-05-29 19:10:48
|
It seems that in the setuptools way [TM] it is rather difficult to not include this folder. Does it work if you only remove those hardcoded paths? It should be easy for me to fix that by not having it search for system libraries when building a tarball. Mike On 05/29/2013 02:55 PM, Christoph Gohlke wrote: > Hi, > > minor issue: the tarball contains a lib\matplotlib.egg-info directory. > The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and > therefore fails to build on my system. Deleting the > lib\matplotlib.egg-info folder works. > > Does the release announcement mention that Windows binaries won't > include required dependencies (pytz, dateutil, pyparsing, six) any > longer? > > Christoph > > > On 5/29/2013 11:41 AM, Michael Droettboom wrote: >> I'm pleased to announce the tagging of matplotlib-1.3.0rc1. >> >> Once the binaries from Christoph and Russell have been uploaded, I'll >> make a broader announcement to get some testing of this in advance of >> the final release. >> >> The tarball is available here: >> >> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz >> >> >> >> The documentation for this version is viewable here: >> >> http://matplotlib.org/1.3.0 >> >> Thanks everyone for their hard work getting this out the door! >> >> Mike >> |
|
From: Christoph G. <cg...@uc...> - 2013-05-29 19:30:33
|
Removing the hardcoded paths works.
I noticed two more build issues.
1) On Python 3.x I see the following errors/warnings, however the build
succeeds.
```
Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\artist.py
line 1265: ParseError: bad token: type=55, value="'", context=('',
(1265, 38))
Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py line
3344: TokenError: ('EOF in multi-line statement', (2, 0))
Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py line
3345: ParseError: bad input: type=5, value=' ',
context=('\n\n<snip>\n\n', (3345, 0))
Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py line
3346: ParseError: bad input: type=5, value=' ',
context=('\n\n<snip>\n\n', (3346, 0))
Can't parse docstring in
build\lib.win-amd64-3.2\matplotlib\backends\backend_agg.py line 322:
TokenError: ('EOF in multi-line statement', (2, 0))
```
2) On Python 2.6 with distribute-0.6.45, the build fails during
packaging. See <https://github.com/matplotlib/matplotlib/pull/2085>
Christoph
On 5/29/2013 12:08 PM, Michael Droettboom wrote:
> It seems that in the setuptools way [TM] it is rather difficult to not
> include this folder. Does it work if you only remove those hardcoded
> paths? It should be easy for me to fix that by not having it search for
> system libraries when building a tarball.
>
> Mike
>
> On 05/29/2013 02:55 PM, Christoph Gohlke wrote:
>> Hi,
>>
>> minor issue: the tarball contains a lib\matplotlib.egg-info directory.
>> The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and
>> therefore fails to build on my system. Deleting the
>> lib\matplotlib.egg-info folder works.
>>
>> Does the release announcement mention that Windows binaries won't
>> include required dependencies (pytz, dateutil, pyparsing, six) any
>> longer?
>>
>> Christoph
>>
>>
>> On 5/29/2013 11:41 AM, Michael Droettboom wrote:
>>> I'm pleased to announce the tagging of matplotlib-1.3.0rc1.
>>>
>>> Once the binaries from Christoph and Russell have been uploaded, I'll
>>> make a broader announcement to get some testing of this in advance of
>>> the final release.
>>>
>>> The tarball is available here:
>>>
>>> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz
>>>
>>>
>>>
>>> The documentation for this version is viewable here:
>>>
>>> http://matplotlib.org/1.3.0
>>>
>>> Thanks everyone for their hard work getting this out the door!
>>>
>>> Mike
>>>
>
|
|
From: Michael D. <md...@st...> - 2013-05-29 19:42:23
|
On 05/29/2013 03:30 PM, Christoph Gohlke wrote:
> Removing the hardcoded paths works.
>
> I noticed two more build issues.
>
> 1) On Python 3.x I see the following errors/warnings, however the
> build succeeds.
>
> ```
> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\artist.py
> line 1265: ParseError: bad token: type=55, value="'", context=('',
> (1265, 38))
> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py
> line 3344: TokenError: ('EOF in multi-line statement', (2, 0))
> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py
> line 3345: ParseError: bad input: type=5, value=' ',
> context=('\n\n<snip>\n\n', (3345, 0))
> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py
> line 3346: ParseError: bad input: type=5, value=' ',
> context=('\n\n<snip>\n\n', (3346, 0))
> Can't parse docstring in
> build\lib.win-amd64-3.2\matplotlib\backends\backend_agg.py line 322:
> TokenError: ('EOF in multi-line statement', (2, 0))
> ```
While annoying, I think these are benign --- the 2to3 tool is tripping
up on these docstrings. Overtime, the docstrings will get rewritten in
numpydoc format and we *might* see these go away.
>
> 2) On Python 2.6 with distribute-0.6.45, the build fails during
> packaging. See <https://github.com/matplotlib/matplotlib/pull/2085>
Good catch. Thanks.
Mike
>
> Christoph
>
>
> On 5/29/2013 12:08 PM, Michael Droettboom wrote:
>> It seems that in the setuptools way [TM] it is rather difficult to not
>> include this folder. Does it work if you only remove those hardcoded
>> paths? It should be easy for me to fix that by not having it search for
>> system libraries when building a tarball.
>>
>> Mike
>>
>> On 05/29/2013 02:55 PM, Christoph Gohlke wrote:
>>> Hi,
>>>
>>> minor issue: the tarball contains a lib\matplotlib.egg-info directory.
>>> The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and
>>> therefore fails to build on my system. Deleting the
>>> lib\matplotlib.egg-info folder works.
>>>
>>> Does the release announcement mention that Windows binaries won't
>>> include required dependencies (pytz, dateutil, pyparsing, six) any
>>> longer?
>>>
>>> Christoph
>>>
>>>
>>> On 5/29/2013 11:41 AM, Michael Droettboom wrote:
>>>> I'm pleased to announce the tagging of matplotlib-1.3.0rc1.
>>>>
>>>> Once the binaries from Christoph and Russell have been uploaded, I'll
>>>> make a broader announcement to get some testing of this in advance of
>>>> the final release.
>>>>
>>>> The tarball is available here:
>>>>
>>>> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz
>>>>
>>>>
>>>>
>>>>
>>>> The documentation for this version is viewable here:
>>>>
>>>> http://matplotlib.org/1.3.0
>>>>
>>>> Thanks everyone for their hard work getting this out the door!
>>>>
>>>> Mike
>>>>
>>
|
|
From: Michael D. <md...@st...> - 2013-05-29 19:50:41
|
Ok, folks. 1.3.0rc2 is here: https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc2/matplotlib-1.3.0rc2.tar.gz Mike On 05/29/2013 03:41 PM, Michael Droettboom wrote: > On 05/29/2013 03:30 PM, Christoph Gohlke wrote: >> Removing the hardcoded paths works. >> >> I noticed two more build issues. >> >> 1) On Python 3.x I see the following errors/warnings, however the >> build succeeds. >> >> ``` >> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\artist.py >> line 1265: ParseError: bad token: type=55, value="'", context=('', >> (1265, 38)) >> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >> line 3344: TokenError: ('EOF in multi-line statement', (2, 0)) >> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >> line 3345: ParseError: bad input: type=5, value=' ', >> context=('\n\n<snip>\n\n', (3345, 0)) >> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >> line 3346: ParseError: bad input: type=5, value=' ', >> context=('\n\n<snip>\n\n', (3346, 0)) >> Can't parse docstring in >> build\lib.win-amd64-3.2\matplotlib\backends\backend_agg.py line 322: >> TokenError: ('EOF in multi-line statement', (2, 0)) >> ``` > While annoying, I think these are benign --- the 2to3 tool is tripping > up on these docstrings. Overtime, the docstrings will get rewritten in > numpydoc format and we *might* see these go away. > >> 2) On Python 2.6 with distribute-0.6.45, the build fails during >> packaging. See <https://github.com/matplotlib/matplotlib/pull/2085> > Good catch. Thanks. > > Mike >> Christoph >> >> >> On 5/29/2013 12:08 PM, Michael Droettboom wrote: >>> It seems that in the setuptools way [TM] it is rather difficult to not >>> include this folder. Does it work if you only remove those hardcoded >>> paths? It should be easy for me to fix that by not having it search for >>> system libraries when building a tarball. >>> >>> Mike >>> >>> On 05/29/2013 02:55 PM, Christoph Gohlke wrote: >>>> Hi, >>>> >>>> minor issue: the tarball contains a lib\matplotlib.egg-info directory. >>>> The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and >>>> therefore fails to build on my system. Deleting the >>>> lib\matplotlib.egg-info folder works. >>>> >>>> Does the release announcement mention that Windows binaries won't >>>> include required dependencies (pytz, dateutil, pyparsing, six) any >>>> longer? >>>> >>>> Christoph >>>> >>>> >>>> On 5/29/2013 11:41 AM, Michael Droettboom wrote: >>>>> I'm pleased to announce the tagging of matplotlib-1.3.0rc1. >>>>> >>>>> Once the binaries from Christoph and Russell have been uploaded, I'll >>>>> make a broader announcement to get some testing of this in advance of >>>>> the final release. >>>>> >>>>> The tarball is available here: >>>>> >>>>> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz >>>>> >>>>> >>>>> >>>>> >>>>> The documentation for this version is viewable here: >>>>> >>>>> http://matplotlib.org/1.3.0 >>>>> >>>>> Thanks everyone for their hard work getting this out the door! >>>>> >>>>> Mike >>>>> > > ------------------------------------------------------------------------------ > Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET > Get 100% visibility into your production application - at no cost. > Code-level diagnostics for performance bottlenecks with <2% overhead > Download for free and get started troubleshooting in minutes. > http://p.sf.net/sfu/appdyn_d2d_ap1 > _______________________________________________ > Matplotlib-devel mailing list > Mat...@li... > https://lists.sourceforge.net/lists/listinfo/matplotlib-devel |
|
From: Christoph G. <cg...@uc...> - 2013-05-29 20:12:55
|
Sorry, but that tarball does not include setup.py and other files. It is only ~25 MB. Christoph On 5/29/2013 12:48 PM, Michael Droettboom wrote: > Ok, folks. 1.3.0rc2 is here: > > https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc2/matplotlib-1.3.0rc2.tar.gz > > > Mike > > > > On 05/29/2013 03:41 PM, Michael Droettboom wrote: >> On 05/29/2013 03:30 PM, Christoph Gohlke wrote: >>> Removing the hardcoded paths works. >>> >>> I noticed two more build issues. >>> >>> 1) On Python 3.x I see the following errors/warnings, however the >>> build succeeds. >>> >>> ``` >>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\artist.py >>> line 1265: ParseError: bad token: type=55, value="'", context=('', >>> (1265, 38)) >>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>> line 3344: TokenError: ('EOF in multi-line statement', (2, 0)) >>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>> line 3345: ParseError: bad input: type=5, value=' ', >>> context=('\n\n<snip>\n\n', (3345, 0)) >>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>> line 3346: ParseError: bad input: type=5, value=' ', >>> context=('\n\n<snip>\n\n', (3346, 0)) >>> Can't parse docstring in >>> build\lib.win-amd64-3.2\matplotlib\backends\backend_agg.py line 322: >>> TokenError: ('EOF in multi-line statement', (2, 0)) >>> ``` >> While annoying, I think these are benign --- the 2to3 tool is tripping >> up on these docstrings. Overtime, the docstrings will get rewritten in >> numpydoc format and we *might* see these go away. >> >>> 2) On Python 2.6 with distribute-0.6.45, the build fails during >>> packaging. See <https://github.com/matplotlib/matplotlib/pull/2085> >> Good catch. Thanks. >> >> Mike >>> Christoph >>> >>> >>> On 5/29/2013 12:08 PM, Michael Droettboom wrote: >>>> It seems that in the setuptools way [TM] it is rather difficult to not >>>> include this folder. Does it work if you only remove those hardcoded >>>> paths? It should be easy for me to fix that by not having it search for >>>> system libraries when building a tarball. >>>> >>>> Mike >>>> >>>> On 05/29/2013 02:55 PM, Christoph Gohlke wrote: >>>>> Hi, >>>>> >>>>> minor issue: the tarball contains a lib\matplotlib.egg-info directory. >>>>> The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and >>>>> therefore fails to build on my system. Deleting the >>>>> lib\matplotlib.egg-info folder works. >>>>> >>>>> Does the release announcement mention that Windows binaries won't >>>>> include required dependencies (pytz, dateutil, pyparsing, six) any >>>>> longer? >>>>> >>>>> Christoph >>>>> >>>>> >>>>> On 5/29/2013 11:41 AM, Michael Droettboom wrote: >>>>>> I'm pleased to announce the tagging of matplotlib-1.3.0rc1. >>>>>> >>>>>> Once the binaries from Christoph and Russell have been uploaded, I'll >>>>>> make a broader announcement to get some testing of this in advance of >>>>>> the final release. >>>>>> >>>>>> The tarball is available here: >>>>>> >>>>>> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> The documentation for this version is viewable here: >>>>>> >>>>>> http://matplotlib.org/1.3.0 >>>>>> >>>>>> Thanks everyone for their hard work getting this out the door! >>>>>> >>>>>> Mike >>>>>> >> >> ------------------------------------------------------------------------------ >> Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET >> Get 100% visibility into your production application - at no cost. >> Code-level diagnostics for performance bottlenecks with <2% overhead >> Download for free and get started troubleshooting in minutes. >> http://p.sf.net/sfu/appdyn_d2d_ap1 >> _______________________________________________ >> Matplotlib-devel mailing list >> Mat...@li... >> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel > > > ------------------------------------------------------------------------------ > Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET > Get 100% visibility into your production application - at no cost. > Code-level diagnostics for performance bottlenecks with <2% overhead > Download for free and get started troubleshooting in minutes. > http://p.sf.net/sfu/appdyn_d2d_ap1 > _______________________________________________ > Matplotlib-devel mailing list > Mat...@li... > https://lists.sourceforge.net/lists/listinfo/matplotlib-devel > > |
|
From: Michael D. <md...@st...> - 2013-05-30 00:41:25
|
Must be a problem with the upload. I'm reuploading it now. I'm not going to bother upping the version -- it's just the result of a failed transfer apparently. Mike On 05/29/2013 04:12 PM, Christoph Gohlke wrote: > Sorry, but that tarball does not include setup.py and other files. It is > only ~25 MB. > > Christoph > > > On 5/29/2013 12:48 PM, Michael Droettboom wrote: >> Ok, folks. 1.3.0rc2 is here: >> >> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc2/matplotlib-1.3.0rc2.tar.gz >> >> >> Mike >> >> >> >> On 05/29/2013 03:41 PM, Michael Droettboom wrote: >>> On 05/29/2013 03:30 PM, Christoph Gohlke wrote: >>>> Removing the hardcoded paths works. >>>> >>>> I noticed two more build issues. >>>> >>>> 1) On Python 3.x I see the following errors/warnings, however the >>>> build succeeds. >>>> >>>> ``` >>>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\artist.py >>>> line 1265: ParseError: bad token: type=55, value="'", context=('', >>>> (1265, 38)) >>>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>>> line 3344: TokenError: ('EOF in multi-line statement', (2, 0)) >>>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>>> line 3345: ParseError: bad input: type=5, value=' ', >>>> context=('\n\n<snip>\n\n', (3345, 0)) >>>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>>> line 3346: ParseError: bad input: type=5, value=' ', >>>> context=('\n\n<snip>\n\n', (3346, 0)) >>>> Can't parse docstring in >>>> build\lib.win-amd64-3.2\matplotlib\backends\backend_agg.py line 322: >>>> TokenError: ('EOF in multi-line statement', (2, 0)) >>>> ``` >>> While annoying, I think these are benign --- the 2to3 tool is tripping >>> up on these docstrings. Overtime, the docstrings will get rewritten in >>> numpydoc format and we *might* see these go away. >>> >>>> 2) On Python 2.6 with distribute-0.6.45, the build fails during >>>> packaging. See <https://github.com/matplotlib/matplotlib/pull/2085> >>> Good catch. Thanks. >>> >>> Mike >>>> Christoph >>>> >>>> >>>> On 5/29/2013 12:08 PM, Michael Droettboom wrote: >>>>> It seems that in the setuptools way [TM] it is rather difficult to not >>>>> include this folder. Does it work if you only remove those hardcoded >>>>> paths? It should be easy for me to fix that by not having it search for >>>>> system libraries when building a tarball. >>>>> >>>>> Mike >>>>> >>>>> On 05/29/2013 02:55 PM, Christoph Gohlke wrote: >>>>>> Hi, >>>>>> >>>>>> minor issue: the tarball contains a lib\matplotlib.egg-info directory. >>>>>> The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and >>>>>> therefore fails to build on my system. Deleting the >>>>>> lib\matplotlib.egg-info folder works. >>>>>> >>>>>> Does the release announcement mention that Windows binaries won't >>>>>> include required dependencies (pytz, dateutil, pyparsing, six) any >>>>>> longer? >>>>>> >>>>>> Christoph >>>>>> >>>>>> >>>>>> On 5/29/2013 11:41 AM, Michael Droettboom wrote: >>>>>>> I'm pleased to announce the tagging of matplotlib-1.3.0rc1. >>>>>>> >>>>>>> Once the binaries from Christoph and Russell have been uploaded, I'll >>>>>>> make a broader announcement to get some testing of this in advance of >>>>>>> the final release. >>>>>>> >>>>>>> The tarball is available here: >>>>>>> >>>>>>> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> The documentation for this version is viewable here: >>>>>>> >>>>>>> http://matplotlib.org/1.3.0 >>>>>>> >>>>>>> Thanks everyone for their hard work getting this out the door! >>>>>>> >>>>>>> Mike >>>>>>> >>> ------------------------------------------------------------------------------ >>> Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET >>> Get 100% visibility into your production application - at no cost. >>> Code-level diagnostics for performance bottlenecks with <2% overhead >>> Download for free and get started troubleshooting in minutes. >>> http://p.sf.net/sfu/appdyn_d2d_ap1 >>> _______________________________________________ >>> Matplotlib-devel mailing list >>> Mat...@li... >>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel >> >> ------------------------------------------------------------------------------ >> Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET >> Get 100% visibility into your production application - at no cost. >> Code-level diagnostics for performance bottlenecks with <2% overhead >> Download for free and get started troubleshooting in minutes. >> http://p.sf.net/sfu/appdyn_d2d_ap1 >> _______________________________________________ >> Matplotlib-devel mailing list >> Mat...@li... >> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel >> >> > ------------------------------------------------------------------------------ > Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET > Get 100% visibility into your production application - at no cost. > Code-level diagnostics for performance bottlenecks with <2% overhead > Download for free and get started troubleshooting in minutes. > http://p.sf.net/sfu/appdyn_d2d_ap1 > _______________________________________________ > Matplotlib-devel mailing list > Mat...@li... > https://lists.sourceforge.net/lists/listinfo/matplotlib-devel |
|
From: Christoph G. <cg...@uc...> - 2013-05-30 02:25:02
|
OK, that worked. I uploaded Windows binaries, which include all test files. Christoph On 5/29/2013 5:34 PM, Michael Droettboom wrote: > Must be a problem with the upload. I'm reuploading it now. I'm not > going to bother upping the version -- it's just the result of a failed > transfer apparently. > > Mike > > On 05/29/2013 04:12 PM, Christoph Gohlke wrote: >> Sorry, but that tarball does not include setup.py and other files. It is >> only ~25 MB. >> >> Christoph >> >> >> On 5/29/2013 12:48 PM, Michael Droettboom wrote: >>> Ok, folks. 1.3.0rc2 is here: >>> >>> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc2/matplotlib-1.3.0rc2.tar.gz >>> >>> >>> Mike >>> >>> >>> >>> On 05/29/2013 03:41 PM, Michael Droettboom wrote: >>>> On 05/29/2013 03:30 PM, Christoph Gohlke wrote: >>>>> Removing the hardcoded paths works. >>>>> >>>>> I noticed two more build issues. >>>>> >>>>> 1) On Python 3.x I see the following errors/warnings, however the >>>>> build succeeds. >>>>> >>>>> ``` >>>>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\artist.py >>>>> line 1265: ParseError: bad token: type=55, value="'", context=('', >>>>> (1265, 38)) >>>>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>>>> line 3344: TokenError: ('EOF in multi-line statement', (2, 0)) >>>>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>>>> line 3345: ParseError: bad input: type=5, value=' ', >>>>> context=('\n\n<snip>\n\n', (3345, 0)) >>>>> Can't parse docstring in build\lib.win-amd64-3.2\matplotlib\axes.py >>>>> line 3346: ParseError: bad input: type=5, value=' ', >>>>> context=('\n\n<snip>\n\n', (3346, 0)) >>>>> Can't parse docstring in >>>>> build\lib.win-amd64-3.2\matplotlib\backends\backend_agg.py line 322: >>>>> TokenError: ('EOF in multi-line statement', (2, 0)) >>>>> ``` >>>> While annoying, I think these are benign --- the 2to3 tool is tripping >>>> up on these docstrings. Overtime, the docstrings will get rewritten in >>>> numpydoc format and we *might* see these go away. >>>> >>>>> 2) On Python 2.6 with distribute-0.6.45, the build fails during >>>>> packaging. See <https://github.com/matplotlib/matplotlib/pull/2085> >>>> Good catch. Thanks. >>>> >>>> Mike >>>>> Christoph >>>>> >>>>> >>>>> On 5/29/2013 12:08 PM, Michael Droettboom wrote: >>>>>> It seems that in the setuptools way [TM] it is rather difficult to not >>>>>> include this folder. Does it work if you only remove those hardcoded >>>>>> paths? It should be easy for me to fix that by not having it search for >>>>>> system libraries when building a tarball. >>>>>> >>>>>> Mike >>>>>> >>>>>> On 05/29/2013 02:55 PM, Christoph Gohlke wrote: >>>>>>> Hi, >>>>>>> >>>>>>> minor issue: the tarball contains a lib\matplotlib.egg-info directory. >>>>>>> The SOURCES.txt contains absolute paths (/usr/src/CXX/...) and >>>>>>> therefore fails to build on my system. Deleting the >>>>>>> lib\matplotlib.egg-info folder works. >>>>>>> >>>>>>> Does the release announcement mention that Windows binaries won't >>>>>>> include required dependencies (pytz, dateutil, pyparsing, six) any >>>>>>> longer? >>>>>>> >>>>>>> Christoph >>>>>>> >>>>>>> >>>>>>> On 5/29/2013 11:41 AM, Michael Droettboom wrote: >>>>>>>> I'm pleased to announce the tagging of matplotlib-1.3.0rc1. >>>>>>>> >>>>>>>> Once the binaries from Christoph and Russell have been uploaded, I'll >>>>>>>> make a broader announcement to get some testing of this in advance of >>>>>>>> the final release. >>>>>>>> >>>>>>>> The tarball is available here: >>>>>>>> >>>>>>>> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3.0rc1/matplotlib-1.3.0rc1.tar.gz >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> The documentation for this version is viewable here: >>>>>>>> >>>>>>>> http://matplotlib.org/1.3.0 >>>>>>>> >>>>>>>> Thanks everyone for their hard work getting this out the door! >>>>>>>> >>>>>>>> Mike >>>>>>>> >>>> ------------------------------------------------------------------------------ >>>> Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET >>>> Get 100% visibility into your production application - at no cost. >>>> Code-level diagnostics for performance bottlenecks with <2% overhead >>>> Download for free and get started troubleshooting in minutes. >>>> http://p.sf.net/sfu/appdyn_d2d_ap1 >>>> _______________________________________________ >>>> Matplotlib-devel mailing list >>>> Mat...@li... >>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel >>> >>> ------------------------------------------------------------------------------ >>> Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET >>> Get 100% visibility into your production application - at no cost. >>> Code-level diagnostics for performance bottlenecks with <2% overhead >>> Download for free and get started troubleshooting in minutes. >>> http://p.sf.net/sfu/appdyn_d2d_ap1 >>> _______________________________________________ >>> Matplotlib-devel mailing list >>> Mat...@li... >>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel >>> >>> >> ------------------------------------------------------------------------------ >> Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET >> Get 100% visibility into your production application - at no cost. >> Code-level diagnostics for performance bottlenecks with <2% overhead >> Download for free and get started troubleshooting in minutes. >> http://p.sf.net/sfu/appdyn_d2d_ap1 >> _______________________________________________ >> Matplotlib-devel mailing list >> Mat...@li... >> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel > > > ------------------------------------------------------------------------------ > Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET > Get 100% visibility into your production application - at no cost. > Code-level diagnostics for performance bottlenecks with <2% overhead > Download for free and get started troubleshooting in minutes. > http://p.sf.net/sfu/appdyn_d2d_ap1 > _______________________________________________ > Matplotlib-devel mailing list > Mat...@li... > https://lists.sourceforge.net/lists/listinfo/matplotlib-devel > > > |
|
From: Russell E. O. <ro...@uw...> - 2013-05-29 21:19:46
|
In article <51A...@st...>, Michael Droettboom <md...@st...> wrote: > I'm pleased to announce the tagging of matplotlib-1.3.0rc1. > > Once the binaries from Christoph and Russell have been uploaded, I'll > make a broader announcement to get some testing of this in advance of > the final release. > > The tarball is available here: > > https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3 > .0rc1/matplotlib-1.3.0rc1.tar.gz > > The documentation for this version is viewable here: > > http://matplotlib.org/1.3.0 > > Thanks everyone for their hard work getting this out the door! It looks like the ability to include pytz and other dependencies in binary distributions has been removed? Is this really what we want? In the past we always included them. Excluding them certainly makes sense if using pip or a similar installer that can handle dependencies, but that is not the case for binaries. I guess we could serve the associated packages (pytz, dateutil and six), or if they can be installed by pip, ask users to install those. But users using binary installers may not even have pip available, so it's a big initial hurdle. I'd personally be happier reverting to the old system. Sorry if I missed a discussion on this. -- Russell |
|
From: Chris B. - N. F. <chr...@no...> - 2013-05-29 23:02:52
|
On Wed, May 29, 2013 at 2:19 PM, Russell E. Owen <ro...@uw...> wrote: > I guess we could serve the associated packages (pytz, dateutil and six), > or if they can be installed by pip, ask users to install those. But > users using binary installers may not even have pip available, so it's a > big initial hurdle. If the binary installers are available (and easy to find), not such a big deal -- this is teh case with Christoph's repository for Windows, for instance. Russell, have you been following the thread I started on the pythonmac list? We really need a way to deal better with binaries on the Mac, including dependency handling. Note that supposedly the "wheel" format is coming (soon?), and after that support for binary wheels by pip. Of course, none of that helps right now... -Chris -- Christopher Barker, Ph.D. Oceanographer Emergency Response Division NOAA/NOS/OR&R (206) 526-6959 voice 7600 Sand Point Way NE (206) 526-6329 fax Seattle, WA 98115 (206) 526-6317 main reception Chr...@no... |
|
From: Russell O. <ro...@uw...> - 2013-05-29 23:56:16
|
On May 29, 2013, at 4:01 PM, Chris Barker - NOAA Federal <chr...@no...> wrote: > On Wed, May 29, 2013 at 2:19 PM, Russell E. Owen <ro...@uw...> wrote: > >> I guess we could serve the associated packages (pytz, dateutil and six), >> or if they can be installed by pip, ask users to install those. But >> users using binary installers may not even have pip available, so it's a >> big initial hurdle. > > If the binary installers are available (and easy to find), not such a > big deal -- this is teh case with Christoph's repository for Windows, > for instance. > > Russell, have you been following the thread I started on the pythonmac > list? We really need a way to deal better with binaries on the Mac, > including dependency handling. > > Note that supposedly the "wheel" format is coming (soon?), and after > that support for binary wheels by pip. > > Of course, none of that helps right now... I have been following that thread with great interest. If such a repository becomes available I do intend to contribute to it. For the record: - I prefer static libraries - The wheel format sounds very promising. But until we have it and until pip supports it, my suspicion is that it's not worth trying to automate dependency handling, and we should live with mpkg binary installers. In the short term I'd be happier if matplotlib continued to support including dateutils, pytz and six. It is not ideal to include dependencies, but convenient and pragmatic. Failing that, i think we should ask users to install pip and use pip to install dateutils, pytz and six. I fear it would be too much work to provide binary installers for pytz, dateutils and six because we would need to make different packages for different versions of python (which feels silly for pure python packages!), keep them up to date, find somewhere to serve them... In the long term, I hope wheel format will allow us to use pip to install matplotlib, in which case it will handle dependencies and we need not include them. -- Russell |
|
From: Michael D. <md...@st...> - 2013-05-30 00:58:53
|
On 05/29/2013 05:19 PM, Russell E. Owen wrote: > In article <51A...@st...>, > Michael Droettboom <md...@st...> > wrote: > >> I'm pleased to announce the tagging of matplotlib-1.3.0rc1. >> >> Once the binaries from Christoph and Russell have been uploaded, I'll >> make a broader announcement to get some testing of this in advance of >> the final release. >> >> The tarball is available here: >> >> https://downloads.sourceforge.net/project/matplotlib/matplotlib/matplotlib-1.3 >> .0rc1/matplotlib-1.3.0rc1.tar.gz >> >> The documentation for this version is viewable here: >> >> http://matplotlib.org/1.3.0 >> >> Thanks everyone for their hard work getting this out the door! > It looks like the ability to include pytz and other dependencies in > binary distributions has been removed? It's really just that the matplotlib source no longer includes them. Binaries can be built however we want them to be. Not knowing how the .dmg is put together, is it possible to add pytz, dateutil, pyparsing and six to the dmg during its creation? > > Is this really what we want? In the past we always included them. > Excluding them certainly makes sense if using pip or a similar installer > that can handle dependencies, but that is not the case for binaries. Right -- and we're not using pip (because we can't) to handle our C/C++ dependencies, many of which we continue to ship with matplotlib. > > I guess we could serve the associated packages (pytz, dateutil and six), > or if they can be installed by pip, ask users to install those. But > users using binary installers may not even have pip available, so it's a > big initial hurdle. > > I'd personally be happier reverting to the old system. Sorry if I missed > a discussion on this. > No worries. I'm certainly guilty of not always being able to keep up with the firehose myself. This work was done as part of MEP11 ( https://github.com/matplotlib/matplotlib/wiki/MEP11) <https://github.com/matplotlib/matplotlib/issues/2092>, and searching the mailing list with "MEP11" as a key should help you find the rest of the discussion. It was pretty unanimous to move to not including these dependencies -- I certainly understand that different communities and platforms have different expectations, though, and if we want to find a way to include the dependencies in the Mac installer, I'm not opposed to that, but I would like to find a way to do it that only affects that one particular use case and not all the others (i.e. something external to or at least optional within the current setup infrastructure). Mike |
|
From: Chris B. - N. F. <chr...@no...> - 2013-05-30 16:26:14
|
On Wed, May 29, 2013 at 5:54 PM, Michael Droettboom <md...@st...> wrote: > It looks like the ability to include pytz and other dependencies in > binary distributions has been removed? > > It's really just that the matplotlib source no longer includes them. > Binaries can be built however we want them to be. Not knowing how the .dmg > is put together, is it possible to add pytz, dateutil, pyparsing and six to > the dmg during its creation? I agree -- adding them to the binary package is the way to go here -- it's packaging issue, not a development or building issue. I can't imagine it would be hard to write a little script that includes them all in one mpkg. > Right -- and we're not using pip (because we can't) to handle our C/C++ > dependencies, many of which we continue to ship with matplotlib. Should the code that's shipped with MPL be build-able with PIP? But the harder issue is third-party dependencies that we expect the system to provide: libpng, libfreetype, .... On the building side, I'd really like to see more support for this -- the Linux package managers handle it OK on LInux, but there is no good system for Windows or OS-X. I'm taking a look at gattai: http://sourceforge.net/projects/gattai/ Mostly for the Mac, but it does support Windows and linux as well. I'm not looking at Windows right now, as Christoph's repository gives me all I need -- which makes me think: Christoph, do you have a build system for all those that might be a good basis for a multi-platform solution? -Chris -- Christopher Barker, Ph.D. Oceanographer Emergency Response Division NOAA/NOS/OR&R (206) 526-6959 voice 7600 Sand Point Way NE (206) 526-6329 fax Seattle, WA 98115 (206) 526-6317 main reception Chr...@no... |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X