scipy 0.10.0 release schedule update
Hi all, For those of you who are wondering what the status of the 0.10.0 release is, we should have RC1 ready very soon. The blocker was/is that single precision Arpack routines on OS X are completely broken. For now the plan is to simply switch to double precision. See https://github.com/scipy/scipy/pull/98 There was one other unresolved issue that came up for the beta release, which is that on Windows there's a Cython bug which prevents compilation with any released Cython version. We can't switch to the latest released Cython + bugfix, because the io.matlab Cython files can't be compiled with them: http://projects.scipy.org/scipy/ticket/1524. My assessment of this issue is that it's not a blocker for the release. Please speak up if you disagree. Cheers, Ralf
Hi, On Sun, Oct 30, 2011 at 12:41 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
Hi all,
For those of you who are wondering what the status of the 0.10.0 release is, we should have RC1 ready very soon. The blocker was/is that single precision Arpack routines on OS X are completely broken. For now the plan is to simply switch to double precision. See https://github.com/scipy/scipy/pull/98
There was one other unresolved issue that came up for the beta release, which is that on Windows there's a Cython bug which prevents compilation with any released Cython version. We can't switch to the latest released Cython + bugfix, because the io.matlab Cython files can't be compiled with them: http://projects.scipy.org/scipy/ticket/1524. My assessment of this issue is that it's not a blocker for the release. Please speak up if you disagree.
I've got some other work I'd like to do on that (matlab io) code - so if you can wait a few days, I will try and fix that. See you, Matthew
On Sun, Oct 30, 2011 at 9:04 PM, Matthew Brett <matthew.brett@gmail.com>wrote:
Hi,
Hi all,
For those of you who are wondering what the status of the 0.10.0 release is, we should have RC1 ready very soon. The blocker was/is that single
On Sun, Oct 30, 2011 at 12:41 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote: precision
Arpack routines on OS X are completely broken. For now the plan is to simply switch to double precision. See https://github.com/scipy/scipy/pull/98
There was one other unresolved issue that came up for the beta release, which is that on Windows there's a Cython bug which prevents compilation with any released Cython version. We can't switch to the latest released Cython + bugfix, because the io.matlab Cython files can't be compiled with them: http://projects.scipy.org/scipy/ticket/1524. My assessment of this issue is that it's not a blocker for the release. Please speak up if you disagree.
I've got some other work I'd like to do on that (matlab io) code - so if you can wait a few days, I will try and fix that.
Sure, that would be great.
Ralf
Hi, On Sun, Oct 30, 2011 at 2:31 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
Ralf
It's just a one-liner fix to the compilation for cython 0.15, as far as I can see: diff --git a/scipy/io/matlab/mio5_utils.pyx b/scipy/io/matlab/mio5_utils.pyx index cb15a00..1604a61 100644 --- a/scipy/io/matlab/mio5_utils.pyx +++ b/scipy/io/matlab/mio5_utils.pyx @@ -169,7 +169,7 @@ cdef class VarReader5: * mat_dtype (bool) * squeeze_me (bool) """ - def __new__(self, preader): + def __cinit__(self, preader): byte_order = preader.byte_order self.is_swapped = byte_order == swapped_code if self.is_swapped: Do you want me to build the c files with Cython 0.15 as well? I've also put in a pull request to add the nice 'cythonize' script by Dag Sverre: https://github.com/scipy/scipy/pull/99 It's useful for making sure the c files are up to date. See you, Matthew
On Sun, Oct 30, 2011 at 11:22 PM, Matthew Brett <matthew.brett@gmail.com>wrote:
Hi,
On Sun, Oct 30, 2011 at 2:31 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
Ralf
It's just a one-liner fix to the compilation for cython 0.15, as far as I can see:
That's good news.
diff --git a/scipy/io/matlab/mio5_utils.pyx b/scipy/io/matlab/mio5_utils.pyx index cb15a00..1604a61 100644 --- a/scipy/io/matlab/mio5_utils.pyx +++ b/scipy/io/matlab/mio5_utils.pyx @@ -169,7 +169,7 @@ cdef class VarReader5: * mat_dtype (bool) * squeeze_me (bool) """ - def __new__(self, preader): + def __cinit__(self, preader): byte_order = preader.byte_order self.is_swapped = byte_order == swapped_code if self.is_swapped:
Do you want me to build the c files with Cython 0.15 as well?
Either way is fine. All Cython files should be regenerated in any case with this fix included: https://github.com/cython/cython/commit/0443ad3d55f0a4762d4009bc606cb98ee4f4... Thanks, Ralf
I've also put in a pull request to add the nice 'cythonize' script by Dag Sverre:
https://github.com/scipy/scipy/pull/99
It's useful for making sure the c files are up to date.
Hi, On Sun, Oct 30, 2011 at 3:56 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
On Sun, Oct 30, 2011 at 11:22 PM, Matthew Brett <matthew.brett@gmail.com> wrote:
Hi,
On Sun, Oct 30, 2011 at 2:31 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
Ralf
It's just a one-liner fix to the compilation for cython 0.15, as far as I can see:
That's good news.
diff --git a/scipy/io/matlab/mio5_utils.pyx b/scipy/io/matlab/mio5_utils.pyx index cb15a00..1604a61 100644 --- a/scipy/io/matlab/mio5_utils.pyx +++ b/scipy/io/matlab/mio5_utils.pyx @@ -169,7 +169,7 @@ cdef class VarReader5: * mat_dtype (bool) * squeeze_me (bool) """ - def __new__(self, preader): + def __cinit__(self, preader): byte_order = preader.byte_order self.is_swapped = byte_order == swapped_code if self.is_swapped:
Do you want me to build the c files with Cython 0.15 as well?
Either way is fine. All Cython files should be regenerated in any case with this fix included: https://github.com/cython/cython/commit/0443ad3d55f0a4762d4009bc606cb98ee4f4...
Maybe I'll leave the building to you then - so we can make sure they all get built right? I'll just submit the thing above as a tiny pull request. as I believe it has no functional consequences. See you, Matthew
On Mon, Oct 31, 2011 at 2:46 AM, Matthew Brett <matthew.brett@gmail.com> wrote:
Hi,
On Sun, Oct 30, 2011 at 3:56 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
On Sun, Oct 30, 2011 at 11:22 PM, Matthew Brett <matthew.brett@gmail.com> wrote:
Hi,
On Sun, Oct 30, 2011 at 2:31 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
Ralf
It's just a one-liner fix to the compilation for cython 0.15, as far as I can see:
That's good news.
diff --git a/scipy/io/matlab/mio5_utils.pyx b/scipy/io/matlab/mio5_utils.pyx index cb15a00..1604a61 100644 --- a/scipy/io/matlab/mio5_utils.pyx +++ b/scipy/io/matlab/mio5_utils.pyx @@ -169,7 +169,7 @@ cdef class VarReader5: * mat_dtype (bool) * squeeze_me (bool) """ - def __new__(self, preader): + def __cinit__(self, preader): byte_order = preader.byte_order self.is_swapped = byte_order == swapped_code if self.is_swapped:
Do you want me to build the c files with Cython 0.15 as well?
Either way is fine. All Cython files should be regenerated in any case with this fix included: https://github.com/cython/cython/commit/0443ad3d55f0a4762d4009bc606cb98ee4f4...
Maybe I'll leave the building to you then - so we can make sure they all get built right?
I'll just submit the thing above as a tiny pull request. as I believe it has no functional consequences.
Ralf, I finally prepared a pull request for fixing the distribution fit() with fixed parameters, including tests. Since the current version is broken it would be good to get this into 0.10. Thanks, Josef
See you,
Matthew _______________________________________________ SciPy-Dev mailing list SciPy-Dev@scipy.org http://mail.scipy.org/mailman/listinfo/scipy-dev
On Mon, Oct 31, 2011 at 7:46 AM, Matthew Brett <matthew.brett@gmail.com>wrote:
Hi,
On Sun, Oct 30, 2011 at 3:56 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
On Sun, Oct 30, 2011 at 11:22 PM, Matthew Brett <matthew.brett@gmail.com
wrote:
Hi,
On Sun, Oct 30, 2011 at 2:31 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
Ralf
It's just a one-liner fix to the compilation for cython 0.15, as far as I can see:
That's good news.
diff --git a/scipy/io/matlab/mio5_utils.pyx b/scipy/io/matlab/mio5_utils.pyx index cb15a00..1604a61 100644 --- a/scipy/io/matlab/mio5_utils.pyx +++ b/scipy/io/matlab/mio5_utils.pyx @@ -169,7 +169,7 @@ cdef class VarReader5: * mat_dtype (bool) * squeeze_me (bool) """ - def __new__(self, preader): + def __cinit__(self, preader): byte_order = preader.byte_order self.is_swapped = byte_order == swapped_code if self.is_swapped:
Do you want me to build the c files with Cython 0.15 as well?
Either way is fine. All Cython files should be regenerated in any case
with
this fix included:
https://github.com/cython/cython/commit/0443ad3d55f0a4762d4009bc606cb98ee4f4...
Maybe I'll leave the building to you then - so we can make sure they all get built right?
I'll just submit the thing above as a tiny pull request. as I believe it has no functional consequences.
Sounds good.
Ralf
On Mon, Oct 31, 2011 at 2:25 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
On Mon, Oct 31, 2011 at 7:46 AM, Matthew Brett <matthew.brett@gmail.com> wrote:
Hi,
On Sun, Oct 30, 2011 at 3:56 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
On Sun, Oct 30, 2011 at 11:22 PM, Matthew Brett <matthew.brett@gmail.com> wrote:
Hi,
On Sun, Oct 30, 2011 at 2:31 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
Ralf
It's just a one-liner fix to the compilation for cython 0.15, as far as I can see:
That's good news.
diff --git a/scipy/io/matlab/mio5_utils.pyx b/scipy/io/matlab/mio5_utils.pyx index cb15a00..1604a61 100644 --- a/scipy/io/matlab/mio5_utils.pyx +++ b/scipy/io/matlab/mio5_utils.pyx @@ -169,7 +169,7 @@ cdef class VarReader5: * mat_dtype (bool) * squeeze_me (bool) """ - def __new__(self, preader): + def __cinit__(self, preader): byte_order = preader.byte_order self.is_swapped = byte_order == swapped_code if self.is_swapped:
Do you want me to build the c files with Cython 0.15 as well?
Either way is fine. All Cython files should be regenerated in any case with this fix included:
https://github.com/cython/cython/commit/0443ad3d55f0a4762d4009bc606cb98ee4f4...
Maybe I'll leave the building to you then - so we can make sure they all get built right?
I'll just submit the thing above as a tiny pull request. as I believe it has no functional consequences.
Sounds good.
What happened to the bfgs endless loop? I don't find a pull request or information anymore. It would be nice to have it in the next release. Josef
Ralf
_______________________________________________ SciPy-Dev mailing list SciPy-Dev@scipy.org http://mail.scipy.org/mailman/listinfo/scipy-dev
On Tue, Nov 1, 2011 at 2:07 AM, <josef.pktd@gmail.com> wrote:
On Mon, Oct 31, 2011 at 2:25 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
On Mon, Oct 31, 2011 at 7:46 AM, Matthew Brett <matthew.brett@gmail.com> wrote:
Hi,
On Sun, Oct 30, 2011 at 3:56 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
On Sun, Oct 30, 2011 at 11:22 PM, Matthew Brett <matthew.brett@gmail.com> wrote:
Hi,
On Sun, Oct 30, 2011 at 2:31 PM, Ralf Gommers <ralf.gommers@googlemail.com> wrote:
Ralf
It's just a one-liner fix to the compilation for cython 0.15, as far as I can see:
That's good news.
diff --git a/scipy/io/matlab/mio5_utils.pyx b/scipy/io/matlab/mio5_utils.pyx index cb15a00..1604a61 100644 --- a/scipy/io/matlab/mio5_utils.pyx +++ b/scipy/io/matlab/mio5_utils.pyx @@ -169,7 +169,7 @@ cdef class VarReader5: * mat_dtype (bool) * squeeze_me (bool) """ - def __new__(self, preader): + def __cinit__(self, preader): byte_order = preader.byte_order self.is_swapped = byte_order == swapped_code if self.is_swapped:
Do you want me to build the c files with Cython 0.15 as well?
Either way is fine. All Cython files should be regenerated in any case with this fix included:
https://github.com/cython/cython/commit/0443ad3d55f0a4762d4009bc606cb98ee4f4...
Maybe I'll leave the building to you then - so we can make sure they all get built right?
I'll just submit the thing above as a tiny pull request. as I believe it has no functional consequences.
Sounds good.
What happened to the bfgs endless loop? I don't find a pull request or information anymore.
https://github.com/scipy/scipy/pull/96
It would be nice to have it in the next release.
That's fine with me.
Ralf
participants (3)
-
josef.pktd＠gmail.com -
Matthew Brett -
Ralf Gommers