sage-6.3-x86_64-Darwin-OSX_10.6_x86_64-app.dmg won't be built?

[Daniel Friedan]

For OS X 10.6, there is no -app build of Sage 6.3 available for download, only
      sage-6.3-x86_64-Darwin-OSX_10.9_x86_64-app.dmg
      sage-6.2-x86_64-Darwin-OSX_10.6_x86_64-app.dmg

Is this a new policy?  Will those of us who don't like OS X >10.6 have to learn to build the OS X 10.6 -app versions ourselves?

thanks,
Daniel

[Volker Braun]

Maybe you want to review this ticket: http://trac.sagemath.org/ticket/16796

[kcrisman]

Maybe you want to review this ticket: http://trac.sagemath.org/ticket/16796

What Volker means by this is that he doesn't have access to a 10.6 buildbot to try this out.  Apparently the machine that was there gave up the ghost... ?  (Volker, did anyone get back to you on the status of bsd?)

 

For OS X 10.6, there is no -app build of Sage 6.3 available for download, only
      sage-6.3-x86_64-Darwin-OSX_10.9_x86_64-app.dmg
      sage-6.2-x86_64-Darwin-OSX_10.6_x86_64-app.dmg

Is this a new policy?  Will those of us who don't like OS X >10.6 have to learn to build the OS X 10.6 -app versions ourselves?

Hopefully not, but it may require volunteers to create them.  It might even be possible for you to make your own version of it, true.  I have a 10.7 binary which I have offered to upload but of course that wouldn't help you.

- kcrisman

[Volker Braun]

On Monday, August 18, 2014 10:45:02 PM UTC+1, kcrisman wrote:

Maybe you want to review this ticket: http://trac.sagemath.org/ticket/16796

What Volker means by this is that he doesn't have access to a 10.6 buildbot to try this out.  Apparently the machine that was there gave up the ghost... ?  (Volker, did anyone get back to you on the status of bsd?)

No, I don't know what the status of bsd is.

[William A Stein]

It's an old desktop in my office.  It may have hung due to heat (I have a fan blowing at it and a window open, but...).  I will check today.

 

--
William Stein
Professor of Mathematics
University of Washington
http://wstein.org
wst...@uw.edu

[Daniel Friedan]

I didn't mean to sound whiny.  I'm very appreciative of all the work being done to make Sage useful.  I'll adapt if the OS X 10.6 app version stops being distributed.

thanks,
Daniel

[William A Stein]

On Wed, Aug 20, 2014 at 9:45 PM, Daniel Friedan <dfri...@gmail.com> wrote:
> I didn't mean to sound whiny. I'm very appreciative of all the work being
> done to make Sage useful. I'll adapt if the OS X 10.6 app version stops
> being distributed.

My office computer died, and that was the only 10.6 machine we had
access to. So I think we have to discontinue 10.6 support, unless
somebody has a better idea.

-- William

>
> thanks,
> Daniel
>
>
> On Monday, August 18, 2014 2:26:27 PM UTC, Daniel Friedan wrote:
>>
>> For OS X 10.6, there is no -app build of Sage 6.3 available for download,
>> only
>> sage-6.3-x86_64-Darwin-OSX_10.9_x86_64-app.dmg
>> sage-6.2-x86_64-Darwin-OSX_10.6_x86_64-app.dmg
>>
>> Is this a new policy? Will those of us who don't like OS X >10.6 have to
>> learn to build the OS X 10.6 -app versions ourselves?
>>
>> thanks,
>> Daniel
>>

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

[kcrisman]

On Thursday, August 21, 2014 7:55:05 AM UTC-4, Stein William wrote:

On Wed, Aug 20, 2014 at 9:45 PM, Daniel Friedan <dfri...@gmail.com> wrote:
> I didn't mean to sound whiny.  I'm very appreciative of all the work being
> done to make Sage useful.  I'll adapt if the OS X 10.6 app version stops
> being distributed.

My office computer died, and that was the only 10.6 machine we had
access to.  So I think we have to discontinue 10.6 support, unless
somebody has a better idea.

One thing to think about - is there a buildbot process for sourceforge-hosted projects?  CRAN (for R) is pretty regular with its requirements of all packages building on all supported platforms and they seem to have this figured out.  But I don't know if there is a generic thing like this available to OSS projects.

[Paul Mercat]

I have an OS X 10.6 computer and I know how to compile sage. I you tell me how to do I can provide you the binaries of the last version of sage.

[Daniel Friedan]

I also have an OS X 10.6.8 machine I could use to build Sage.  I've never compiled Sage myself, so I'd need some guidance.

Daniel

[Daniel Friedan]

I've done a couple of experiments with the latest version of Sage (6.3) under OS X 10.6.8, using the official 10.7 and 10.6 binaries available for download.

(1) sage-6.3-x86_64-Darwin-OSX_10.7_x86_64.dmg

This is an app version (the file name should be changed to sage-6.3-x86_64-Darwin-OSX_10.7_x86_64-app.dmg).

It runs on OS X 10.6.8 (though I didn't test much).

It seems large, 4.53 GB.

(2) sage-6.3-x86_64-Darwin-OSX_10.6_x86_64.dmg

This is the non-app version.  I turned it into an app version by the following:
      1.  moved sage/ out of sage-6.2.app/Contents/Resources/
      2.  copied sage-6.2.app/ and named the copy sage-6.3.app/
      3.  moved the 6.3 version of sage/ into sage-6.3.app/Contents/Resources/
      4.  edit a few lines in sage-6.3.app/Contents/Info.plist (actually, I should have but didn't bother).

So it seems trivial to make an OS X 10.6 app version from a OS X 10.6 non-app version.

The size is 3.55 GB, smaller by about 1GB than the official OS X 10.7 app version.

For comparison, the official sage-6.2-app is 3.32 GB.

Daniel

[Volker Braun]

git checkout 6.3

export SAGE_APP_DMG=yes

export SAGE_APP_BUNDLE=yes

./sage -bdist

[Dima Pasechnik]

On 2014-08-22, Daniel Friedan <dfri...@gmail.com> wrote:
> ------=_Part_5_1837348027.1408711654748
> Content-Type: text/plain; charset=UTF-8

>
> I've done a couple of experiments with the latest version of Sage (6.3)
> under OS X 10.6.8, using the official 10.7 and 10.6 binaries available for
> download.

Please also pay attention to the CPU type. If you want your binary to run on all
the machines capable of running OSX 10.6.8, you might need to be more careful than just
running Volker's recepy.
Specifically, Core2 Duo chips are pretty much screwed up under OSX 10.6, as their 64-bit
capabilities aren't really well-supported.

At least it would be important to know the CPU type you use exactly.

Just in case,
Dima

>
> (1) *sage-6.3-x86_64-Darwin-OSX_10.7_x86_64.dmg*
><http://boxen.math.washington.edu/home/sagemath/sage-mirror/osx/intel/sage-6.3-x86_64-Darwin-OSX_10.7_x86_64.dmg>

>
> This is an app version (the file name should be changed to

> *sage-6.3-x86_64-Darwin-OSX_10.7_x86_64-app.dmg*
><http://boxen.math.washington.edu/home/sagemath/sage-mirror/osx/intel/sage-6.3-x86_64-Darwin-OSX_10.7_x86_64.dmg>
> ).

>
> It runs on OS X 10.6.8 (though I didn't test much).
>
> It seems large, 4.53 GB.
>

> (2) *sage-6.3-x86_64-Darwin-OSX_10.6_x86_64.dmg*
><http://boxen.math.washington.edu/home/sagemath/sage-mirror/osx/intel/sage-6.3-x86_64-Darwin-OSX_10.6_x86_64.dmg>

[kcrisman]

I've done a couple of experiments with the latest version of Sage (6.3) under OS X 10.6.8, using the official 10.7 and 10.6 binaries available for download.

(1) sage-6.3-x86_64-Darwin-OSX_10.7_x86_64.dmg

This is an app version (the file name should be changed to sage-6.3-x86_64-Darwin-OSX_10.7_x86_64-app.dmg).

Yikes, you are right!  Harald, is this easy to do on the mirrors?

 

It runs on OS X 10.6.8 (though I didn't test much).

It seems large, 4.53 GB.

(2) sage-6.3-x86_64-Darwin-OSX_10.6_x86_64.dmg

This is the non-app version.  I turned it into an app version by the following:
      1.  moved sage/ out of sage-6.2.app/Contents/Resources/
      2.  copied sage-6.2.app/ and named the copy sage-6.3.app/
      3.  moved the 6.3 version of sage/ into sage-6.3.app/Contents/Resources/
      4.  edit a few lines in sage-6.3.app/Contents/Info.plist (actually, I should have but didn't bother).

So it seems trivial to make an OS X 10.6 app version from a OS X 10.6 non-app version.

Yes, that will almost certainly work.  In fact, Ivan has built in the ability to choose any Sage executable you want!  (See the preferences for the app.)  So one can totally just keep the same app and download new executables.  But we haven't advertised this since the app version is supposed to be primarily for people who don't know about such things...

The size is 3.55 GB, smaller by about 1GB than the official OS X 10.7 app version.

I probably just did something wrong in compiling it - it's not "official" in the sense that Volker made it.

[Dima Pasechnik]

On 2014-08-22, Daniel Friedan <dfri...@gmail.com> wrote:

> ------=_Part_1227_548338737.1408715857673
> Content-Type: text/plain; charset=UTF-8
>
> My OS X 10.6.8 machine does have and Intel Core 2 Duo processor.
>
> Volker's recipe did not work. Here's what it produced:
>
> I tried following Volker's instructions for building the app bundle under OS X 10.6.8, but it failed.

this the an unclear error message:

> find: /Volumes/4TB-1/Sage_building/sage/local/var/lib/sage/installed: No such file or directory

that is to say, "no, Sage was not installed, one didn't even try..."

He skipped 'make'.

After

$ git clone git://github.com/sagemath/sage.git
$ cd sage

you need to build sage

by running 'make'. To use 2 cores, do the following

$ export MAKE="make -j2"
$ $MAKE

(better do it overnight, as it will take a while, and you might also need to disable
sleeping on your laptop...)




>
> OS X 10.6.8
>
> $ cd /Volumes/4TB-1
> $ mkdir Sage_building
> $ cd Sage_building/
> $ git clone git://github.com/sagemath/sage.git
> Cloning into 'sage'...
> remote: Counting objects: 246211, done.
> remote: Compressing objects: 100% (675/675), done.
> remote: Total 246211 (delta 1082), reused 992 (delta 779)
> Receiving objects: 100% (246211/246211), 74.73 MiB | 7.38 MiB/s, done.
> Resolving deltas: 100% (172257/172257), done.
> Checking connectivity... done.
> Checking out files: 100% (4220/4220), done.
>
> $ cd sage
> $ git checkout 6.3
> Note: checking out '6.3'.
>
> You are in 'detached HEAD' state. You can look around, make experimental
> changes and commit them, and you can discard any commits you make in this
> state without impacting any branches by performing another checkout.
>
> If you want to create a new branch to retain commits you create, you may
> do so (now or later) by using -b with the checkout command again. Example:
>
> git checkout -b new_branch_name
>
> HEAD is now at 00199fb... Updated Sage version to 6.3
>
> $ export SAGE_APP_DMG=yes
> $ export SAGE_APP_BUNDLE=yes
> $ ./sage -bdist
> find: /Volumes/4TB-1/Sage_building/sage/local/var/lib/sage/installed: No such file or directory
> Sage version 6.3, release date 2014-08-10
> rm -rf config configure build/Makefile-auto.in
> % Total % Received % Xferd Average Speed Time Time Time Current
> Dload Upload Total Spent Left Speed
> 100 89470 100 89470 0 0 105k 0 --:--:-- --:--:-- --:--:-- 215k
> Cloning into '/Volumes/4TB-1/Sage_building/sage/tmp/sage-6.3-i386-Darwin'...
> warning: templates not found /Volumes/4TB-1/Sage_building/sage/local/share/git-core/templates
> done.
> Checking out files: 100% (4220/4220), done.
> Checking connectivity: 246211, done.
> Finished cloning Sage sources
> Copying files over to tmp directory
> tar: local: Cannot stat: No such file or directory
> tar: Error exit delayed from previous errors.
> cp: /Volumes/4TB-1/Sage_building/sage/local/bin/sage-README-osx.txt: No such file or directory
> $
>
>
>
>
> On Friday, August 22, 2014 1:47:15 PM UTC, Dima Pasechnik wrote:

[Daniel Friedan]

I already deleted the posting that Dima is replying to.  I misread the instructions and omit the 'make' step,

[Harald Schilly]

On Fri, Aug 22, 2014 at 4:16 PM, kcrisman <kcri...@gmail.com> wrote:
> Harald, is this easy to do on the mirrors?

Easy and tedious. As long as the filename changes, it counts as being
removed and a new one is added.

-- H

[Daniel Friedan]

Again, trying to build Sage 6.3 under OS X 10.6.8 (Intel Core 2 duo cpu).

This time I did run 'make', but ended with "** BUILD FAILED **"

$ git clone git://github.com/sagemath/sage.git
$ cd sage

$ export MAKE="make -j2"
$ $MAKE

...[wait 6 hours]...
$ git checkout 6.3

$ export SAGE_APP_DMG=yes
$ export SAGE_APP_BUNDLE=yes
$ ./sage -bdist

The tail of the output from './sage -bdist' was

CompileC build/Sage.build/Debug/Sage.build/Objects-normal/i386/AppController.o AppController.m normal i386 objective-c com.apple.compilers.llvmgcc42
    cd /NOT_ON_SSD/Software/sage_build/sage/src/mac-app
    setenv LANG en_US.US-ASCII
    /Developer/usr/bin/llvm-gcc-4.2 -x objective-c -arch i386 -fmessage-length=0 -pipe -std=gnu99 -Wno-trigraphs -fpascal-strings -O0 -Wreturn-type -Wunused-variable -fasm-blocks -mmacosx-version-min=10.4 -gdwarf-2 -fvisibility=hidden -iquote /NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Sage-generated-files.hmap -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Sage-own-target-headers.hmap -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Sage-all-target-headers.hmap -iquote /NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Sage-project-headers.hmap -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Debug/include -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/DerivedSources/i386 -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/DerivedSources -F/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Debug -include /var/folders/3P/3P9Z7q3dGRexXy8O5coyfk+++TI/-Caches-/com.apple.Xcode.501/SharedPrecompiledHeaders/Sage_Prefix-gloqutariauqusgrjhneruvqyalg/Sage_Prefix.pch -c /NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m -o /NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Objects-normal/i386/AppController.o
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m: In function '-[AppController taskTerminated:]':
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:199: error: 'NSModalResponse' undeclared (first use in this function)
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:199: error: (Each undeclared identifier is reported only once
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:199: error: for each function it appears in.)
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:199: error: expected ';' before 'resp'
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:200: error: 'resp' undeclared (first use in this function)
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:200: error: 'NSModalResponseOK' undeclared (first use in this function)
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m: In function '-[AppController ensureReadWrite]':
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:362: error: 'NSModalResponse' undeclared (first use in this function)
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:362: error: expected ';' before 'resp'
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:363: error: 'resp' undeclared (first use in this function)
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:363: error: 'NSModalResponseOK' undeclared (first use in this function)
/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/AppController.m:366: error: 'NSModalResponseCancel' undeclared (first use in this function)

CompileC build/Sage.build/Debug/Sage.build/Objects-normal/i386/PreferencePanelController.o PreferencePanelController.m normal i386 objective-c com.apple.compilers.llvmgcc42
    cd /NOT_ON_SSD/Software/sage_build/sage/src/mac-app
    setenv LANG en_US.US-ASCII
    /Developer/usr/bin/llvm-gcc-4.2 -x objective-c -arch i386 -fmessage-length=0 -pipe -std=gnu99 -Wno-trigraphs -fpascal-strings -O0 -Wreturn-type -Wunused-variable -fasm-blocks -mmacosx-version-min=10.4 -gdwarf-2 -fvisibility=hidden -iquote /NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Sage-generated-files.hmap -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Sage-own-target-headers.hmap -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Sage-all-target-headers.hmap -iquote /NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Sage-project-headers.hmap -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Debug/include -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/DerivedSources/i386 -I/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/DerivedSources -F/NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Debug -include /var/folders/3P/3P9Z7q3dGRexXy8O5coyfk+++TI/-Caches-/com.apple.Xcode.501/SharedPrecompiledHeaders/Sage_Prefix-gloqutariauqusgrjhneruvqyalg/Sage_Prefix.pch -c /NOT_ON_SSD/Software/sage_build/sage/src/mac-app/PreferencePanelController.m -o /NOT_ON_SSD/Software/sage_build/sage/src/mac-app/build/Sage.build/Debug/Sage.build/Objects-normal/i386/PreferencePanelController.o


** BUILD FAILED **


The following build commands failed:
    CompileC build/Sage.build/Debug/Sage.build/Objects-normal/i386/AppController.o AppController.m normal i386 objective-c com.apple.compilers.llvmgcc42
(1 failure)
Failed to build Sage.app.
If you don't wish to build Sage.app set SAGE_APP_BUNDLE=no
$

[Dima Pasechnik]

On 2014-08-22, Daniel Friedan <dfri...@gmail.com> wrote:

> ------=_Part_1632_1626995138.1408746138628
> Content-Type: text/plain; charset=UTF-8

>
> Again, trying to build Sage 6.3 under OS X 10.6.8 (Intel Core 2 duo cpu).
>
> This time I did run 'make', but ended with "** BUILD FAILED **"
>
> $ git clone git://github.com/sagemath/sage.git
> $ cd sage
> $ export MAKE="make -j2"
> $ $MAKE
> ...[wait 6 hours]...
> $ git checkout 6.3
> $ export SAGE_APP_DMG=yes
> $ export SAGE_APP_BUNDLE=yes
> $ ./sage -bdist
>

Actually,
$ git checkout 6.3
should obviously have come before $MAKE, as
this is the version you want to build and distribute
(hopefully this was OK, as

$ git clone git://github.com/sagemath/sage.git

checks out version 6.3 anyway, IMHO)


Anyhow, you seem to hit the bug fixed in http://trac.sagemath.org/16796
(this fix is not in 6.3)

You need to apply it, I suppose.
Have a look at the Sage developer guide for detailed instructions
(this is basically 3 git commands you need, as outlined below)


first of all, you need to add trac git server to the list of remotes:

$ git remote add trac git://trac.sagemath.org/sage.git -t master

then, we get the patch from trac.sagemath.org/16796

$ git fetch trac u/iandrus/trac-16796

finally, we apply it:

$ git checkout FETCH_HEAD

now
./sage -bdist
should succeed
(no need to rebuild Sage, as we only changed src/mac-app/AppController.m
which is not a part of Sage proper)

HTH,
Dima

[Volker Braun]

On Saturday, August 23, 2014 12:08:40 AM UTC+1, Dima Pasechnik wrote:

Anyhow, you seem to hit the bug fixed in http://trac.sagemath.org/16796
(this fix is not in 6.3)

Ah yes, forgot that we are talking about the OSX < 10.9 bug. This will be in 6.4.beta2, so it might be best to wait a week or so until it is ready.

[Daniel Friedan]

Dima's suggestion worked.  My OS X 10.6.8 build of Sage 6.3 succeeded, produced a .dmg file containing an app version that ran and passed some rudimentary tests.

I have a few questions about the resulting build.  Modulo these questions, I could volunteer to produce OSX 10.6 builds of future Sage versions (both app and non-app).  I'd probably need some hand-holding when/if things go wrong.

The .dmg file that was produced was named
       sage-6.3-i386-Darwin.dmg
This differs from the name of the Sage 6.2 .dmg that I downloaded and used previously:
       sage-6.2-x86_64-Darwin-OSX_10.6_x86_64-app.dmg

Did I leave out some build option to get 64bit?
How can I determine if my build is 64bit?
The name of the .dmg file does not contain '-app' or 'OSX_10.6'.  Did I leave out build options?

Here's what I did for the build.  I am using a Core 2 Duo cpu running OS X 10.6.8.

git clone git://github.com/sagemath/sage.git
cd sage
export MAKE="make -j2"
$MAKE

git checkout 6.3

git remote add trac git://trac.sagemath.org/sage.git -t master

git fetch trac u/iandrus/trac-16796
git checkout FETCH_HEAD

export SAGE_APP_DMG=yes
export SAGE_APP_BUNDLE=yes
./sage -bdist

thanks,
Daniel

[William A Stein]

Hi,

I just want to note that the second contributor ever to Sage -- David
Kohel -- was complaining to me yesterday about some sort of lack of
support of Sage right now for his OS X 10.6 laptop. Just a point in
favor of continued support of 10.6.

-- William

[Volker Braun]

You have to manually rename the sage-6.3-i386-Darwin.dmg to include the -app and version details.

It seems you built 32-bit binaries, which you can verify with "file" e.g. On the old 10.6 buildbot the output defaulted to 64 bit. No idea why that is or how to change the defaults for xcode commandline tools.

[Dima Pasechnik]

On 2014-08-26, Volker Braun <vbrau...@gmail.com> wrote:
> ------=_Part_2182_349848151.1409057890862
> Content-Type: text/plain; charset=UTF-8

>
> You have to manually rename the sage-6.3-i386-Darwin.dmg to include the
> -app and version details.
>
> It seems you built 32-bit binaries, which you can verify with "file" e.g.
> On the old 10.6 buildbot the output defaulted to 64 bit. No idea why that
> is or how to change the defaults for xcode commandline tools.

I presume the old buildbot had better hardware than Core2 Duo (that's what
the OP has in a laptop).
For the latter one cannot really get OSX10.6 running in 64-bit mode.
Apple apparently preferred users upgrading to 10.7,
instead of providing 64-bit support for these CPUs.

We have seen these issues few years ago, when these
machines hit the market.
As I have such a laptop (running 64-bit Debian Linux now :-))
myself, I knew this all too well.
(But I can't recall details now; all I recall was quite a
few complaints that binary releases of Sage didn't work)

[Volker Braun]

On Tuesday, August 26, 2014 2:59:37 PM UTC+1, Dima Pasechnik wrote:

Apple apparently preferred users upgrading to 10.7, 

Considering that Apple dropped support for 10.6 it is imho grossly negligent to connect a OSX 10.6 machine to the internet. Upgrade to OSX >=10.7 or Linux ;-)

 

[kcrisman]

Apple apparently preferred users upgrading to 10.7, 

Considering that Apple dropped support for 10.6 it is imho grossly negligent to connect a OSX 10.6 machine to the internet. Upgrade to OSX >=10.7 or Linux ;-)

 

You always say stuff like that ("buy the poor guy a few gig of ram", etc.).    But it's not "grossly" negligent if you aren't going to phishing sites and it's probably not really an option for a lot of users to upgrade or switch to Linux - it is not JUST inertia that has all those people keeping XP or even older versions of Windows, for instance.   As an example, I knew a lot of people who upgraded to 10.5 back in the day from 10.4 and really regretted it because it made the machine so much slower.  Especially regarding Linux conversion, I can throw a couple thick (and scholarly) books I just read about real costs that entail at you while you duck - and I don't say that because I think Linux is bad.  Plop this argument into the developing world with few IT resources of the sort needed to make Linux fun and easy, or to try to just randomly find exactly the chip you need at the price you can pay ... 

[Daniel Friedan]

I've been able to run the 64 bit version of Sage 6.2 onn my Core 2 Duo OS X 10.6 macbookpro with no trouble. 
This is the version I downloaded as
    sage-6.2-x86_64-Darwin-OSX_10.6_x86_64-app.dmg

The problem seems to be with building a 64 bit version.
Is it impossible to build a 64bit version of Sage on this machine?

thanks,
Daniel

[Dima Pasechnik]

On 2014-08-26, Daniel Friedan <dfri...@gmail.com> wrote:
> ------=_Part_4367_1566256233.1409080390865
> Content-Type: text/plain; charset=UTF-8

>
> I've been able to run the 64 bit version of Sage 6.2 onn my Core 2 Duo OS X
> 10.6 macbookpro with no trouble.
> This is the version I downloaded as

> * sage-6.2-x86_64-Darwin-OSX_10.6_x86_64-app.dmg*
><http://boxen.math.washington.edu/home/sagemath/sage-mirror/osx/intel/sage-6.2-x86_64-Darwin-OSX_10.6_x86_64-app.dmg>
>
> The problem seems to be with *building* a 64 bit version.

> Is it impossible to build a 64bit version of Sage on this machine?

Perhaps you just have to

$ export SAGE64=yes

before starting make.

cf. http://www.sagemath.org/doc/installation/source.html
for details.
(it says SAGE64=yes is only needed on OSX 10.4 and 10.5 though)

[Dima Pasechnik]

On 2014-08-26, kcrisman <kcri...@gmail.com> wrote:
> ------=_Part_472_1817143489.1409078230466
> Content-Type: text/plain; charset=UTF-8

>
>
>>
>> Apple apparently preferred users upgrading to 10.7,
>>>
>>
>> Considering that Apple dropped support for 10.6 it is imho grossly
>> negligent to connect a OSX 10.6 machine to the internet. Upgrade to OSX
>> >=10.7 or Linux ;-)
>>
>>
>
> You always say stuff like that ("buy the poor guy a few gig of ram", etc.).
> But it's not "grossly" negligent if you aren't going to phishing sites

well, our sysadmins told me in no uncertain terms few months ago that they
won't let a laptop running OSX 10.6 anywhere near our intranet.
Go figure.

> and it's probably not really an option for a lot of users to upgrade or
> switch to Linux - it is not JUST inertia that has all those people keeping
> XP or even older versions of Windows, for instance. As an example, I knew
> a lot of people who upgraded to 10.5 back in the day from 10.4 and really
> regretted it because it made the machine so much slower. Especially
> regarding Linux conversion, I can throw a couple thick (and scholarly)
> books I just read about real costs that entail at you while you duck - and
> I don't say that because I think Linux is bad. Plop this argument into the
> developing world with few IT resources of the sort needed to make Linux fun
> and easy, or to try to just randomly find exactly the chip you need at the
> price you can pay ...

Developing world is used to run pirated software, don't you known this?
Linux runs on pretty much every Intel box, provided it is not bleeding edge.

Heck, it runs on Apple hardware too - it did so even *before* OSX came about.
The 1st laptop I was able to afford, a toilet seat design iBook, came
with OS 9, and I was running YDL (Yellow Dog Linux) on it.

[Volker Braun]

On Tuesday, August 26, 2014 7:37:10 PM UTC+1, kcrisman wrote:

negligent if you aren't going to phishing sites

Critical bugs could be in every subsystem, not just in the browser. You can't know.

The only thing that is for sure is that, somewhere, somebody is checking every new bugfix that Apple puts out for whether it is exploitable on OSX 10.6.

 

[Daniel Friedan]

Still trying to build 64bit Sage 6.3 on an OS X 10.6.8 Core 2 Duo machine.

Dima's suggestion '$ export SAGE64=yes ' did not work. 

'make' failed -- one package failed to build (error message copied below). 

Note that 'make'  previously succeeded without '$ export SAGE64=yes ' .

'$ ./sage -bdist' succeeded, but produced a 32 bit Sage-6.3.app  which was only 1.1G large (compared to 3.5G for the usual Sage-6.3.app).

=================================================================
2014-08-27 build Sage-6.3 trying for 64bit version

$ git clone git://github.com/sagemath/sage.git
$ cd sage

$ git checkout 6.3
$ export SAGE64=yes

$ export MAKE="make -j2"
$ $MAKE

[... lots of output ...]

***************************************************************
Error building Sage.

The following package(s) may have failed to build:

package: polybori-0.8.3
log file: //NOT_ON_SSD/Software/sage_build/build2/sage/logs/pkgs/polybori-0.8.3.log
build directory: //NOT_ON_SSD/Software/sage_build/build2/sage/local/var/tmp/sage/build/polybori-0.8.3

The build directory may contain configuration files and other potentially
helpful information. WARNING: if you now run 'make' again, the build
directory will, by default, be deleted. Set the environment variable
SAGE_KEEP_BUILT_SPKGS to 'yes' to prevent this.

make: *** [build] Error 1

$ git checkout 6.3

$ git remote add trac git://trac.sagemath.org/sage.git -t master

$ git fetch trac u/iandrus/trac-16796

$ git checkout FETCH_HEAD

$ export SAGE_APP_DMG=yes
$ export SAGE_APP_BUNDLE=yes
$ ./sage -bdist

[Build succeeded but produced 32 bit app version, size only 1.1G]
=================================================================
thanks,
Daniel Friedan

[Dima Pasechnik]

On 2014-08-27, Daniel Friedan <dfri...@gmail.com> wrote:
> ------=_Part_3092_594188982.1409156737747
> Content-Type: text/plain; charset=UTF-8

>
> Still trying to build 64bit Sage 6.3 on an OS X 10.6.8 Core 2 Duo machine.
>
> Dima's suggestion '$ export SAGE64=yes ' did not work.

did it have any effect on the type of executables/libraries that did build OK?
(you can use 'file' in SAGE_LOCAL/lib and SAGE_LOCAL/bin to find out...)

if yes, please post the log related to the package (from SAGE_ROOT/logs/pkgs/)

>
> 'make' failed -- one package failed to build (error message copied below).
>
> Note that 'make' previously succeeded without '$ export SAGE64=yes ' .
>
> '$ ./sage -bdist' succeeded, but produced a 32 bit Sage-6.3.app which was
> only 1.1G large (compared to 3.5G for the usual Sage-6.3.app).

probably something incomplete and not working.

[kcrisman]

>
> You always say stuff like that ("buy the poor guy a few gig of ram", etc.).
>    But it's not "grossly" negligent if you aren't going to phishing sites

well, our sysadmins told me in no uncertain terms few months ago that they
won't let a laptop running OSX 10.6 anywhere near our intranet.
Go figure.

They're paid to worry.  I'm not saying it can't happen, and it does all the time.  But it can also happen with up-to-date (Heartbleed, anyone?).

 

> I don't say that because I think Linux is bad.  Plop this argument into the
> developing world with few IT resources of the sort needed to make Linux fun
> and easy, or to try to just randomly find exactly the chip you need at the
> price you can pay ...

> Developing world is used to run pirated software, don't you known this? 

Well, yes.  But that doesn't make my argument go away, though it was interesting that piracy rates weren't really correlated with open source adoption either direction.

http://mitpress.mit.edu/books/comingled-code

> Linux runs on pretty much every Intel box, provided it is not bleeding edge. 

Interesting article I read about Peace Corps volunteers needing to figure out what to do with machines with < 256 MB of RAM, where Linux wouldn't play nice.

Anyway, this is all OT, sorry.

[Daniel Friedan]

All the executables/libraries are 64bit.  (per output from $ file sage/local/lib/*; file sage/local/bin/*)

Attached is the log for the package that failed: polybori-0.8.3.log

Daniel

[Dima Pasechnik]

On 2014-08-27, Daniel Friedan <dfri...@gmail.com> wrote:

> ------=_Part_4855_138320293.1409165879953
> Content-Type: multipart/alternative;
> boundary="----=_Part_4856_1171858397.1409165879953"
>
> ------=_Part_4856_1171858397.1409165879953
> Content-Type: text/plain; charset=UTF-8

>
> All the executables/libraries are 64bit. (per output from $ file
> sage/local/lib/*; file sage/local/bin/*)
>
> Attached is the log for the package that failed: polybori-0.8.3.log

weird, it ends up with an error in creating a manpage...:

scons: *** [/NOT_ON_SSD/Software/sage_build/build2/sage/local/man/man1/PolyGUI2.1] Source `/NOT_ON_SSD/Software/sage_build/build2/sage/local/man/man1/PolyGUI.1' not found, needed by target `/NOT_ON_SSD/Software/sage_build/build2/sage/local/man/man1/PolyGUI2.1'.
scons: building terminated because of errors.
Error installing PolyBoRi.

What kind of a filesystem do you have mounted on /NOT_ON_SSD/ ?

Could you post the output of

$ mount


A wild guess might be that if it's some kind of vfat thing, then it might misbehave when
you use a 64-bit tool, scons, there...
(or you're just running out of space on the device...)

If it's the case, you might try your build on a proper OSX filesystem instead.

------------------------------

PS. just to make sure you can reproduce it, can you run
$ ./sage -f polybori

in the directory you build Sage in, and see if you end up with the same error?

[Daniel Friedan]

(1) The file system is fine.  Despite the name, '/NOT_ON_SSD/' is just a subdirectory of '/' on the imac's internal disk, which is an SSD with a Journaled HFS+ file system with 140GB free. 
$ mount
/dev/disk0s2 on / (hfs, local, journaled, noatime)

(2) '$ ./sage -f polybori' succeeded:
....
Done installing PolyBoRi.

real    2m12.677s
user    1m59.620s
sys    0m11.988s
Successfully installed polybori-0.8.3
Deleting temporary build directory
/NOT_ON_SSD/Software/sage_build/build2/sage/local/var/tmp/sage/build/polybori-0.8.3
Finished installing polybori-0.8.3.spkg

Warning: it might be needed to update the Sage library before
installed packages work: you should run 'make' from $SAGE_ROOT
before running Sage.

(3) I then tried (without running 'make distclean'):

$ export SAGE64=yes
$ export MAKE="make -j2"
$ $MAKE

and got the same error:

Error building Sage.
The following package(s) may have failed to build:
package: polybori-0.8.3

(4) I'm now trying 'make' again from scratch.  It will take 4-6 hours, depending on whether the error recurs.

Daniel

[Dima Pasechnik]

On Thursday, August 28, 2014 1:03:04 PM UTC+1, Daniel Friedan wrote:

(1) The file system is fine.  Despite the name, '/NOT_ON_SSD/' is just a subdirectory of '/' on the imac's internal disk, which is an SSD with a Journaled HFS+ file system with 140GB free. 
$ mount
/dev/disk0s2 on / (hfs, local, journaled, noatime)

(2) '$ ./sage -f polybori' succeeded:
....
Done installing PolyBoRi.

real    2m12.677s
user    1m59.620s
sys    0m11.988s
Successfully installed polybori-0.8.3
Deleting temporary build directory
/NOT_ON_SSD/Software/sage_build/build2/sage/local/var/tmp/sage/build/polybori-0.8.3
Finished installing polybori-0.8.3.spkg

Warning: it might be needed to update the Sage library before
installed packages work: you should run 'make' from $SAGE_ROOT
before running Sage.

Oh, sorry, I guess ./sage -f might still build it 32-bit.

You need to do

$ SAGE64=yes  ./sage -f polybori

instead.

 

(3) I then tried (without running 'make distclean'):
$ export SAGE64=yes
$ export MAKE="make -j2"
$ $MAKE

and got the same error:
Error building Sage.
The following package(s) may have failed to build:
package: polybori-0.8.3

at that point I would have tried 'make' without 'j2', i.e. building using just one core.

It could be a race condition in the parallel build that causes this.

Next time, when your build ends up with this error, just

do

$ make