[sage-devel] Re: MSI packages for Windows and atlas
Hi All, I am currently in Europe for inteviews, and not able to connect my laptop to the net and access the LinBox svn server. I plan to apply Michael's patches to the svn, and create the new linbox spkg asap right after that, say next monday or so. As for ATLAS, a PIIISSE2 is fine for most Intel32, but athlon64 users will not be happy. i don't see any easy answer for this pb. Maybe just options 3) 4-ATLAS-INTEL32) 4-ATLAS-ATHLON64) ? Sorry for the delay. Clement On May 9, 7:24 am, William Stein [EMAIL PROTECTED] wrote: On Tuesday 08 May 2007 10:19 pm, mabshoff wrote: Yep, I got that. But SAGE's linbox is still months behind the svn version of linbox right now, and that needs to be addressed first. Volunteers? E.g., SAGE uses Integers_GMP, but Linbox deprecated that type... Not sure about the Integers_GMP, but I could ping the Linbox people via linbox-use and ask if there is a canonicalwork-around/ migration path. Basically somebody needs to just get linbox_wrap.cpp to compile with the new version of linbox. I put an updated linbox package here: http://www.sagemath.org/packages/experimental/ It's probably just a matter of changing GMP_Integers to PID_Integer or PID_integer or something like that (I asked Clement Pernet recently). Then build everything, discover bugs, fix, write to linbox mailing lists, etc... :-) Be really happy with how fast the result is. The patch I send you should apply to the version of Linbox in SAGE because it is only a two line patch to blas.m4. The patch isn't even in Linbox svn yet (as of 1.1.3-r2701). From a quick look at numpy and gsl it should be just as easy to insert a couple lines in the right place. Oh, OK, that should be easy. --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: SAGE-2.5 (!)
On May 9, 2007, at 1:06 PM, Hamptonio wrote: Hi, I have the following 'make test' failure for sage 2.5 on a PPC OS X powerbook: The following tests failed: sage -t devel/sage-main/sage/libs/cf/cf.pyx Total time for all tests: 4614.8 seconds Ummm I just did a clean install of sage 2.5 on a powerpc OS X, and the cf directory doesn't even exist. david --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups sage-devel group. To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel?hl=en -~--~~~~--~~--~--~---
[sage-devel] Re: SAGE-2.5 (!)
On May 8, 2007, at 9:10 PM, William Stein wrote: Hello, SAGE-2.5 has been released! Yeah! Great! The new stuff seems to be working well. SAGE is to the point where it can be a valuable tool for a lot of people (other than developers :-}). I built 2.5 from scratch for Mac OS X (10.4.9) on both PowerPC (G5) and Intel (Core Duo). On both, the build completed without complaint. On Intel, 'make test' completed without error (in 1842 seconds). On PowerPC/G5, most tests completed successfully. The ones that didn't seem to be dsage/twisted issues (I haven't looked at the details). That part of the log is appended. Thoughts? Justin === This script will run the unit tests for DSage Running 52 tests. sage.dsage.database.tests.test_clientdb ClientDatabaseTestCase testadd_user ... [ OK] testdel_user ... [ OK] testset_enabled ...[ OK] testupdate_login_time ... [ OK] sage.dsage.database.tests.test_job JobTestCase testcreate_job ... [ OK] testjobCreationTime ...[ OK] testjobFailures ...[ OK] testjobFile ...[ OK] testjobFinishTime ... [ OK] testjobKilled ... [ OK] testjobResult ... [ OK] testjobStatus ... [ OK] testjobUpdateTime ... [ OK] testjob_id ... [ OK] sage.dsage.database.tests.test_jobdb DatabasePrunerTestCase testclean_failed_jobs ... [ OK] testclean_old_jobs ... [ OK] JobDatabaseSQLiteTestCase testcreate_jdict ... [ OK] testget_job ...[ OK] testget_job_by_id ... [ OK] testget_killed_jobs_list ... [ OK] testhas_job ...[ OK] teststore_job ... [ OK] JobDatabaseZODBTestCase testget_active_jobs ...[ OK] testget_job ...[ OK] testget_jobs_by_username ... [ OK] testget_jobs_list ... [ OK] testget_next_job_id ...[ OK] testhas_job ...[ OK] testremove_job ... [ OK] teststore_job ... [ OK] sage.dsage.server.tests.test_server DSageServerTestCase testget_active_clients_list ...[ OK] testget_active_jobs ...[ OK] testget_all_jobs ... [ OK] testget_job ...[ OK] testget_job_by_id ... [ OK] testget_job_result_by_id ... [ OK] testget_jobs_by_username ... [ OK] testget_killed_jobs_list ... [ OK] testjob_done ... [ OK] testjob_failed ... [ OK] testkill_job ... [ OK] testsubmit_job ... [ OK] sage.dsage.twisted.tests.test_pubkeyauth PublicKeyCredentialsCheckerTest testBadLogin ... [ OK] testBadLogin2 ... [ OK] testBadLogin3 ... [ OK]
[sage-devel] SAGE 2.5 DMG images
Hello folks, as promised I did build DMG images for MacOSX. The can be found at http://fsmath.mathematik.uni-dortmund.de/~mabshoff/sage/ 2.5.0 ppc is there, IntelMac is coming right up. There are still some limitations: - When you start the app the Icon in the taskbar keeps jumping - When you close the browser sage itself keeps running - When you restart the app while sage is still running bad things happen - But: You can start an new browser and point to localhost:8000 to work with the old instance - the icon is smushed, I am to lazy to correct that one now - and I am no artist and there is no gimp or anything installed around here on the boxen I have access to (maybe I should chnge that) What to do in the future: - automate the build process - rewrite the wrapper script to check if a sage instance is running, if it is just start a browser - sage should kill itself after a given timeout Feel free to implement any of those features :) In order to clean up after you quiet the browser just kill the appropriate python job. If no other python jobs are running just do a killall python, otherwise do something along the line of ps -aux | grep python and kill the right ones :) Feedback welcome. The MacIntel DMG will show up shortly. I cannot test that one (the IntelMac have already gone home with their owners and I am still waiting on my new laptop to arrive) William, feel free to throw a copy on sage.math in the download directory. My bandwith is free and quick here, so you can download from here as you like. Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: SAGE 2.5 DMG images
Very nice. I'm assuming the associated notebook just lives in the .app bundle? I think SAGE already quits with an error when you try and run the notebook on an already running notebook directory. While none of these issues are that big of a deal to the OS X command line crowd, I would imagine people whose expertise is limited to drag- n-drop could be quite confused and would not know how to/be scared to fix things. Perhaps for that reason we should mark the .dmg distributions as alpha (but hopefully not for long). - Robert On May 9, 2007, at 10:55 AM, mabshoff wrote: There are still some limitations: - When you start the app the Icon in the taskbar keeps jumping - When you close the browser sage itself keeps running - When you restart the app while sage is still running bad things happen - But: You can start an new browser and point to localhost:8000 to work with the old instance - the icon is smushed, I am to lazy to correct that one now - and I am no artist and there is no gimp or anything installed around here on the boxen I have access to (maybe I should chnge that) What to do in the future: - automate the build process - rewrite the wrapper script to check if a sage instance is running, if it is just start a browser - sage should kill itself after a given timeout Feel free to implement any of those features :) In order to clean up after you quiet the browser just kill the appropriate python job. If no other python jobs are running just do a killall python, otherwise do something along the line of ps -aux | grep python and kill the right ones :) Feedback welcome. The MacIntel DMG will show up shortly. I cannot test that one (the IntelMac have already gone home with their owners and I am still waiting on my new laptop to arrive) William, feel free to throw a copy on sage.math in the download directory. My bandwith is free and quick here, so you can download from here as you like. Cheers, Michael --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: SAGE-2.5 (!)
Yes, I was upgrading. -MH On May 9, 12:14 pm, David Harvey [EMAIL PROTECTED] wrote: On May 9, 2007, at 1:06 PM, Hamptonio wrote: Hi, I have the following 'make test' failure for sage 2.5 on a PPC OS X powerbook: The following tests failed: sage -t devel/sage-main/sage/libs/cf/cf.pyx Total time for all tests: 4614.8 seconds Ummm I just did a clean install of sage 2.5 on a powerpc OS X, and the cf directory doesn't even exist. david --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: SAGE-2.5 (!)
On 5/9/07, Justin C. Walker [EMAIL PROTECTED] wrote: On PowerPC/G5, most tests completed successfully. The ones that didn't seem to be dsage/twisted issues (I haven't looked at the details). That part of the log is appended. Thoughts? The dsage developer (Yi) doesn't have access to a ppc mac. He did create a patch for me at the last minute that might have fixed the unit tests, but it was too late for inclusion in sage-2.5. It will certainly be in sage-2.5.1. By the way, in the past make test didn't run the dsage unit tests -- sage-2.5 is the first version to do this, which is cool because it mostly guarantees that dsage will actually work for people on a wider range of systems. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: SAGE-2.5 (!)
On May 9, 2007, at 4:40 PM, didier deshommes wrote: On 5/9/07, William Stein [EMAIL PROTECTED] wrote: * new packages: - flintqs I have not heard about flintqs for awhile and I'm curious about it. Would the flintqs people care to explain how to use it from SAGE? Is it for univariate polynomials over Q only? Will there be a talk about it at SD4 :) ? flintqs is Bill Hart's quadratic sieve code, as packaged for use in SAGE. It's a routine for factoring huge integers. It's already available in SAGE using the qsieve() function. But this has nothing to do with univariate polynomials as such. There is a larger project called FLINT, which has not yet been released. Bill Hart and I are the authors. The quadratic sieve will be a component, when it is released. The other component will be univariate polynomial arithmetic over Z. (Not Q.) Work on the polynomial arithmetic slowed down a while ago because we both got distracted by other things. But recently we have found more time, and development is coming along nicely. I might be persuaded to do a short talk on the current state of FLINT at SD4. I am also considering spending some time at SD4 on starting to develop a patch to wrap the polynomial arithmetic in SAGE. I don't expect it to be done by the end of SD4, but it might be feasible to at least start getting a patch that I can maintain as FLINT matures, until the time is ready to actually bring it into SAGE. David --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: SAGE-2.5 (!)
On May 9, 2007, at 1:07 PM, Yi Qiang wrote: On May 9, 2007, at 10:19 AM, Justin C. Walker wrote: On PowerPC/G5, most tests completed successfully. The ones that didn't seem to be dsage/twisted issues (I haven't looked at the details). That part of the log is appended. Thoughts? Hi Justin, Like William said, I don't have access to a OS X PPC box (maybe you can help out there?) and apparently sysctl reports differently on OS X. Here is a bundle that might fix this on PPC http://sage.math.washington.edu/home/yqiang/dsage.hg Please let me know if this resolves the issue on PPC. Thanks for the patch. I applied this on both my intel and PowerPC systems, and re-ran sage-dsage-trial. 52 tests passed on both systems. If you want me to try things out in the future, feel free to contact me directly. Justin -- Justin C. Walker, Curmudgeon-at-Large () The ASCII Ribbon Campaign /\ Help Cure HTML Email --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] SD4 room-sharing wanted for the first 3 days
Hi there, I am looking for someone to room-share with from Monday 11/06 to Thursday 14/06 for SAGE Days 4. If you would like to split the cost of a room, please email me. I plan to stay at the College Inn ($75/ night). didier --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: SD4 room-sharing wanted for the first 3 days
didier deshommes [EMAIL PROTECTED] writes: Hi there, I am looking for someone to room-share with from Monday 11/06 to Thursday 14/06 for SAGE Days 4. If you would like to split the cost of a room, please email me. I plan to stay at the College Inn ($75/ night). Hi Didier, That might work well for me. William, is there any funding for travel/lodging? Nick --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Python 2.5.1 bug?
I know this isn't a bug in sage, but it is good to be aware of... If the contents of the file crash.py is: file def crash(): n = 5 k = 2 return [( 2**(n-2*i-2) - 2**(floor(n/2)-i-1) - 1 ) for i in range(k)] /file Then you get bizarre behavior: sage: load crash.py sage: crash() Unhandled SIGSEGV: A segmentation fault occured in SAGE. etc So far my ignorance has prevented me from reproducing this in Python alone (I'm too dependent on the sage environment). --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~---
[sage-devel] Re: Python 2.5.1 bug?
On 5/9/07, Robert Miller [EMAIL PROTECTED] wrote: I know this isn't a bug in sage, but it is good to be aware of... If the contents of the file crash.py is: file def crash(): n = 5 k = 2 return [( 2**(n-2*i-2) - 2**(floor(n/2)-i-1) - 1 ) for i in range(k)] /file Then you get bizarre behavior: sage: load crash.py sage: crash() Unhandled SIGSEGV: A segmentation fault occured in SAGE. etc So far my ignorance has prevented me from reproducing this in Python alone (I'm too dependent on the sage environment). Thanks for the bug report! Actually, it's a bug in SAGE-2.5 not Python. It comes from the floor function, which you're implicitly calling, and which is not the same as the Python floor function. In fact, it's a new symbolic function, so cool things like this make sense: sage: f = floor(x^2 + 1) sage: f floor(x^2) + 1 sage: f(x=5) 26 sage: f(x=3/2) 3 sage: show(plot(f)) It calls the floor method on objects if they have them, but didn't correctly deal with the builtin Python numeric types, which don't have a floor method. This led to a problem. Apply the attached patch to fix it, or do hg_sage.pull(), which will also give you a fix to dsage. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~--- 4393.patch Description: Binary data
[sage-devel] Re: Python 2.5.1 bug?
On 5/9/07, William Stein [EMAIL PROTECTED] wrote: I know this isn't a bug in sage, but it is good to be aware of... If the contents of the file crash.py is: file def crash(): n = 5 k = 2 return [( 2**(n-2*i-2) - 2**(floor(n/2)-i-1) - 1 ) for i in range(k)] /file Then you get bizarre behavior: sage: load crash.py sage: crash() Unhandled SIGSEGV: A segmentation fault occured in SAGE. etc The actual SIGSEGV part of the bug was caused by a mistake in the arithmetic architecture code, in which in one case it assumed that an element was of type Element, but it wasn't. This is fixed if you do hg_sage.pull() or apply the attached patch. William --~--~-~--~~~---~--~~ To post to this group, send email to sage-devel@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-devel URLs: http://sage.scipy.org/sage/ and http://modular.math.washington.edu/sage/ -~--~~~~--~~--~--~--- 4386.patch Description: Binary data