Sage 6.9.rc2 released

[Volker Braun]

As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html

698579c Updated Sage version to 6.9.rc2

f99f57d Trac #19324: Force reconfiguration after GCC installation

f603730 Updated Sage version to 6.9.rc1

74721d5 Trac #19125: DeprecatedMethod wrapping is broken

1b5a573 Trac #19294: Doctest failure with coxeter3

2f7c727 Updated Sage version to 6.9.rc0

[kcrisman]

698579c Updated Sage version to 6.9.rc2

f99f57d Trac #19324: Force reconfiguration after GCC installation

f603730 Updated Sage version to 6.9.rc1

74721d5 Trac #19125: DeprecatedMethod wrapping is broken

1b5a573 Trac #19294: Doctest failure with coxeter3

2f7c727 Updated Sage version to 6.9.rc0

If someone can check that the notebook still works properly and so forth that would be nice.  Unfortunately due to some kind of firewall I haven't been able to pull for a while (the sysadmin will look into this on Monday, I think, his first suggestion about proxies didn't work, unfortunately). 

[kcrisman]

(But I'll download the tarball and see what happens too.)

[Rob Beezer]

On SageMathCloud, built from source, passes long tests ("ptestlong").

https://cloud.sagemath.com/projects/53b77207-8614-4086-a032-432af4b4cdbd/files/sage-dev-images/

[Justin C. Walker]

> On Oct 3, 2015, at 07:36, Volker Braun <vbrau...@gmail.com> wrote:
>
> As always, you can get the latest beta version from the "develop" git branch. Alternatively, the self-contained source tarball is at http://www.sagemath.org/download-latest.html

Built from the tarball on OS X, 10.10.5 (Quad-core Core i7). No problems and all tests (‘ptestlong’) passed!

Justin

--
Justin C. Walker
Director
Institute for the Enhancement of the Director's Income
--
Fame is fleeting, but obscurity
just drags on and on. F&E

[Jan Groenewald]

Hi

On 3 October 2015 at 20:49, kcrisman <kcri...@gmail.com> wrote:

If someone can check that the notebook still works properly and so forth that would be nice.  Unfortunately due to some kind of firewall I haven't been able to pull for a while (the sysadmin will look into this on Monday, I think, his first suggestion about proxies didn't work, unfortunately). 

For future reference, opened port 9418, for the git protocol, which I think was the issue.

Regards,

Jan

--

  .~.
  /V\     Jan Groenewald
 /( )\    www.aims.ac.za
 ^^-^^ 

[kcrisman]

If someone can check that the notebook still works properly and so forth that would be nice.  Unfortunately due to some kind of firewall I haven't been able to pull for a while (the sysadmin will look into this on Monday, I think, his first suggestion about proxies didn't work, unfortunately). 

For future reference, opened port 9418, for the git protocol, which I think was the issue.

Regards,

Jan

Yeah, this seems to have worked - thanks Jan! 

[kcrisman]

With SAGE_INSTALL_CCACHE=yes I get (from tarball)

Found local metadata for ccache-3.2.2

Attempting to download package ccache-3.2.2.tar.bz2 from mirrors

Downloading the Sage mirror list

CRITICAL [mirror_list|_refresh:164]: Downloading the mirror list failed

Also, this causes the entire build to fail, which seems somewhat unreasonable - couldn't it just keep building and report that it won't use ccache?

Note that , particularly mysteriously 

The following package(s) may have failed to build (not necessarily

during this run of 'make all'):

<blank line>

That's not good either.  Also, just setting the env var to blank didn't seem to negate it, currently trying with unsetting it.  That's probably my ignorance of bash, though.

I can provide more details later if you would find them interesting.

[kcrisman]

That's not good either.  Also, just setting the env var to blank didn't seem to negate it, currently trying with unsetting it.  That's probably my ignorance of bash, though.

Aagh, it's *still* trying to download it, even though the env var SAGE_INSTALL_CCACHE is now not even showing up in the list!  Also trying to download gcc 4.9.2 (remember, this is from the tarball).  I will reuntar the ball and see what happens. 

[Henri Girard]

I compiled it on ubuntu 15.04 successfully, sage --notebook=ipython worked, but I had no latex displaying, raw one.

Le 05/10/2015 09:46, kcrisman a écrit :

That's not good either.  Also, just setting the env var to blank didn't seem to negate it, currently trying with unsetting it.  That's probably my ignorance of bash, though.

Aagh, it's *still* trying to download it, even though the env var SAGE_INSTALL_CCACHE is now not even showing up in the list!  Also trying to download gcc 4.9.2 (remember, this is from the tarball).  I will reuntar the ball and see what happens. 

--
You received this message because you are subscribed to the Google Groups "sage-release" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-release...@googlegroups.com.
To post to this group, send email to sage-r...@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-release.
For more options, visit https://groups.google.com/d/optout.

[kcrisman]

I compiled it on ubuntu 15.04 successfully, sage --notebook=ipython worked, but I had no latex displaying, raw one.

Interesting.  I was in the meantime able to upgrade my previous installation and sagenb seems to work fine.  Still interested in that ccache thing - I had to reduce the load on my machine somewhat but now should be able to run a fresh build of rc2 from tarball. 

[kcrisman]

From a fresh untarred ball on Mac Lion:

Found local metadata for gcc-4.9.2.p1

Attempting to download package gcc-4.9.2.tar.bz2 from mirrors

Downloading the Sage mirror list

CRITICAL [mirror_list|_refresh:164]: Downloading the mirror list failed

Traceback (most recent call last):

  File "/Users/.../sage-6.9.rc2/build/bin/sage-download-file", line 28, in <module>

    SageDownloadFileApplication().run()

  File "/Users/.../sage-6.9.rc2/build/bin/../sage_bootstrap/cmdline.py", line 193, in run

    tarball.download()

  File "/Users/.../sage-6.9.rc2/build/bin/../sage_bootstrap/tarball.py", line 148, in download

    for mirror in MirrorList():

  File "/Users/.../sage-6.9.rc2/build/bin/../sage_bootstrap/mirror_list.py", line 63, in __init__

    self.mirrors = self._load()

  File "/Users/.../sage-6.9.rc2/build/bin/../sage_bootstrap/mirror_list.py", line 78, in _load

    return ast.literal_eval(mirror_list)

  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/ast.py", line 49, in literal_eval

    node_or_string = parse(node_or_string, mode='eval')

  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/ast.py", line 37, in parse

    return compile(expr, filename, mode, PyCF_ONLY_AST)

  File "<unknown>", line 0

    

    ^

SyntaxError: unexpected EOF while parsing

make[2]: *** [/Users/.../sage-6.9.rc2/local/var/lib/sage/installed/gcc-4.9.2.p1] Error 1

make[1]: *** [all-toolchain] Error 2

real 6m46.434s

user 1m57.608s

sys 1m7.451s

***************************************************************

Error building Sage.

I view this as a blocker bug.  Internet should not be required for Sage to build from scratch from a tarball.  I'll see whether this is also due to some firewall but in the meantime let's get this fixed.

[kcrisman]

Also, I still get

[Jeroen Demeyer]

On 2015-10-05 09:46, kcrisman wrote:
>
> That's not good either. Also, just setting the env var to blank
> didn't seem to negate it, currently trying with unsetting it.
> That's probably my ignorance of bash, though.
>
>
> Aagh, it's *still* trying to download it, even though the env var
> SAGE_INSTALL_CCACHE is now not even showing up in the list! Also trying
> to download gcc 4.9.2 (remember, this is from the tarball).

gcc should be contained in the source tarball. This is fallout from the
build system changes. Fix needs review at

http://trac.sagemath.org/ticket/19347

[kcrisman]

gcc should be contained in the source tarball. This is fallout from the
build system changes. Fix needs review at

What about the blank line with no failed packages?  Clearly a package failed.  So somehow whatever mechanism causes the package to be listed (I used to know what that is, but in the new build system I don't) is not working if a package is not even downloaded.  But that is still a failure in that package - OR there should be some separate message saying that there is a connectivity problem.  Because the final message isn't the connectivity message, it's the message about Sage not building, and even

SyntaxError: unexpected EOF while parsing

isn't exactly something that screams "you have a bad proxy setting", which is what turned out to be the case for me.

[Jeroen Demeyer]

Right. You bug report actually shows 3 independent bugs:
* GCC shouldn't be downloaded in the first place
* The failed download of the mirror_list doesn't give an error message
* The blank line without failed packages

The first one is fixed at #19347.

The third one was fixed at #18731, but the fix got lost somehow(???) and
is fixed again at #19349.

[kcrisman]

Right. You bug report actually shows 3 independent bugs:

Haha! 

[kcrisman]

By the way, all normal tests passed on the build from scratch once I got the proxy issue sorted out, which is nice.

[Eric Gourgoulhon]

Hi,

On Ubuntu 14.04 64bits, the build was OK but when I try
./sage -n jupyter
I get

The IPython notebook requires ssl, even if you do not use
https. Install the openssl development packages in your system and
then rebuild Python (sage -f python2).

Now the openssl development packages (libssl-dev) are already installed on this computer (Jupyter notebook is launched successfully from sage 6.8). I've tried
./sage -pip install pyopenssl
./sage -f python2
Both were successful, but the problem remains: ./sage -n jupyter returns the same error message as above.

Eric.

[Volker Braun]

whats the output of

$ sage -python

>>> from notebook.notebookapp import NotebookApp

[Eric Gourgoulhon]

Le lundi 5 octobre 2015 21:17:43 UTC+2, Volker Braun a écrit :

whats the output of

$ sage -python

>>> from notebook.notebookapp import NotebookApp

It is 

Python 2.7.9 (default, Oct  5 2015, 20:41:58)
[GCC 4.8.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>>

[Eric Gourgoulhon]

Le lundi 5 octobre 2015 21:17:43 UTC+2, Volker Braun a écrit :

whats the output of

$ sage -python

>>> from notebook.notebookapp import NotebookApp

Sorry for the previous truncated message; the output is

$ ./sage -python

Python 2.7.9 (default, Oct  5 2015, 20:41:58)
[GCC 4.8.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.

>>> from notebook.notebookapp import NotebookApp

Traceback (most recent call last):

  File "<stdin>", line 1, in <module>
ImportError: No module named notebook.notebookapp

[Eric Gourgoulhon]

PS: note that with sage 6.8 (where ./sage -n ipython works), I have the same error message:
$ ./sage -python
Python 2.7.9 (default, Aug  1 2015, 11:59:13)

[GCC 4.8.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> from notebook.notebookapp import NotebookApp
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: No module named notebook.notebookapp

[Volker Braun]

You need to install Sage (i.e. run "make")

[Eric Gourgoulhon]

Le lundi 5 octobre 2015 21:45:22 UTC+2, Volker Braun a écrit :

You need to install Sage (i.e. run "make")

I did; it recompiled many things and took a while.
But the problem is still there: ./sage -n jupyter results in the same
error message.

[Volker Braun]

Well why can't you import notebook then?

[Henri Girard]

Hi,
I build sage 6.9.rc2 on ubuntu 15.04 and got it working but then latex was n't working in ipython notebook command ouput was between frame like a table.
By the way a question : Sage 6.9 rc2 can't be otbain with git clone ?
I could get only the tar download. Otherwise with git it have version number sage 6.8, even on develop button.

[Dima Pasechnik]

On Monday, 5 October 2015 18:15:23 UTC-7, HG wrote:

Hi,
I build sage 6.9.rc2 on ubuntu 15.04 and got it working but then latex was n't working in ipython notebook command ouput was between frame like a table.
By the way a question : Sage 6.9 rc2 can't be otbain with git clone ?
I could get only the tar download. Otherwise with git it have version number sage 6.8, even on develop button.

huh?

see https://github.com/sagemath/sage/tree/develop

for Sage 6.9.rc2

[Henri Girard]

Well, the tree/develop gives sage-develop in zip but in git clone it's sage.git, I am compiling it and will see if it's sage 6.9 maybe I used to make a mistake ?
But last time I did it I had sage6.8

[Jeroen Demeyer]

On 2015-10-05 22:56, Eric Gourgoulhon wrote:
> But the problem is still there: ./sage -n jupyter results in the same
> error message.

OK, first verify that you're on the correct Sage version:

$ cat VERSION.txt

This should output something like
Sage version 6.9.rc1, released 2015-10-01

Next, properly build Sage:

$ ./configure && make

Try it again:

$ ./sage --python
>>> from notebook.notebookapp import NotebookApp

[Henri Girard]

I noticed something which blocked me for a lot of time, maybe you know it ?
After using different sage versions, I faced a problem in .local numpy ?
Not knowing why and being blocked with all what I used I decided to
delete .local ! And everything works again...
If it can help
Henri

[kcrisman]

I have a question about the new "required" use of pretty_print in these rcs.  I created this interact in class today.

@interact
def _(a = 3, f = x^2):
    f(x) = f
    D = diff(f,x)
    I = integrate(f,x)
    pretty_print(html("The derivative of ${}$ at $a = {}$ is ${}$".format(latex(f(x)),a, D(a))))
    pretty_print(html("The indefinite integral of ${}$ is ${}+C$".format(latex(f(x)),latex(I(x)))))
    show(plot(f,(x,a-1,a+1))+plot(D(a)*(x-a)+f(a),(x,a-1,a+1),linestyle='--',color='green'),figsize=3)

This looks quite different without the pretty_print in Sage 6.8 (what we are currently using in the lab here at AIMS); namely, the lines directly follow one another, I guess <p> elements or something (view source doesn't seem to get what I want for interacts).  With the Sage 6.8 behavior for pretty_print, we get big huge spaces in between - not optimal at all for presentation.

I won't be able to get to a computer with rc2 for a bit; does the new pretty_print with html do the (better) behavior of just having separate html lines without huge spaces in between?  Otherwise you have to zoom out so much that it's barely worth using.  Or is there some other way to multiline?  Using \n seems to work okay but that is somewhat annoying compared to actually being able to, you know, write multiple lines...

Thanks for any feedback!
- kcrisman

[Henri Girard]

Hi,
I just compiled sage-6.9 rc2 and running x^2 i got this writing
\newcommand{\Bold}[1]{\mathbf{#1}}x^{2} with frame ?
what is the problem ?
regards
Henri

[Harald Schilly]

Where exactly have you written this, what have you started or done prior to that in Sage, and have there been any errors?

--

[Henri Girard]

sorry... After compiling develop.zip (i didn't find git for develop ?) I started sage --notebook=ipython, in the jupyter notebook I wrote the command %display latex (like I do as usual)  and in another command I just tried x*x which answers me x² in latex . But not today with this version. I compiled it already 3 times last week and didn't use it again but I wanted to try sage-6.9 to see if ipython notebook works.
regards
Henri

[Henri Girard]

sorry : I just noted it's in ipython notebook I don't get latex
sage -notebook have latex
I am sorry for the disturbance

Le 06/10/2015 13:23, Harald Schilly a écrit :

[Eric Gourgoulhon]

Hi,

Le mardi 6 octobre 2015 09:43:26 UTC+2, Jeroen Demeyer a écrit : 

OK, first verify that you're on the correct Sage version:

$ cat VERSION.txt

It results in:
Sage version 6.9.rc2, released 2015-10-03

Next, properly build Sage:

$ ./configure && make

I did it.

Try it again:

$ ./sage --python
 >>> from notebook.notebookapp import NotebookApp

Same error message:

 ImportError: No module named notebook.notebookapp

and
./sage -n jupyter
still refuses to launch the notebook in the browser, with the error message:

The IPython notebook requires ssl, even if you do not use
https. Install the openssl development packages in your system and
then rebuild Python (sage -f python2).

Here is the detail of what I did:

My system is Ubuntu 14.04 64 bits with libssl-dev installed.
Some time ago, I did

git clone https://github.com/sagemath/sage.git
cd sage
make

This built sage 6.8 and
./sage -n ipython
worked fine (Jupyter notebook was correctly opened in the browser).

Yesterday, I did *in the same directory*:

git checkout -b develop
git pull origin develop
git branch -v
* develop 698579c Updated Sage version to 6.9.rc2
  master  7eb8510 Updated Sage version to 6.8
make

This built sage 6.9rc2 (except for some error in building the
documentation, but I now I should perform
make doc-clean && make to fix it).

Now ./sage -n ipython no longer works (nor ./sage -n jupyter, which
is equivalent I think), with the error message
"The IPython notebook requires ssl..."
As suggested by the message, I ran

./sage -f python2

It did not helped.
Today, I've did

./configure && make

It did not helped either.

Am I the only one experiencing this with Ubuntu 14.04.

Best wishes,

Eric.

[Henri Girard]

I can't say now as I have 15.04. To get sage-6.9 rc2 I got the zip on the git sage-develop. I tried sage.git (version 6.8 apparently) before but I couldn't compiled it because an error. After compiling sage-develop
I did in the repertory
./bootstrap
./configure
make -j9
After about 2 hours I had a running sage-6.9rc2 and sage --notebook=ipython works, (but I can't get latex just a table with latex command)
sage -notebook work perfectly and I installed sm-install.sh which runs very good (even latex this way).
As I use texmacs I move it in /usr/share/sage with a link. By the way I always keep the sage-6.8 from ppa, but doesn't work always asking for ssl.
I haven't been able to remake it ./sage -f python2 or python or python3, never works ?
But before I was on 14.04 but never got it working either !
Henri

[Volker Braun]

I don't know whats broken in your install but its unlikely that it can be fixed over email.

Run "make distclean && make" for a clean build.

[kcrisman]

This looks quite different without the pretty_print in Sage 6.8 (what we are currently using in the lab here at AIMS); namely, the lines directly follow one another, I guess <p> elements or something (view source doesn't seem to get what I want for interacts).  With the Sage 6.8 behavior for pretty_print, we get big huge spaces in between - not optimal at all for presentation.

Okay, I was finally able to test this and it seems like both of these now work the same (in the way I think is good) in 6.9.rc2.  Thanks. 

[Eric Gourgoulhon]

Hi,

Le mardi 6 octobre 2015 23:49:52 UTC+2, Volker Braun a écrit :

I don't know whats broken in your install but its unlikely that it can be fixed over email.

Run "make distclean && make" for a clean build.

Indeed, this was the thing to do. I've made a fresh install (via git clone) of sage 6.9rc3 and ./sage -n jupyter works !
Sorry for the noise...

Eric.