Re: [sage-devel] Should sage -i install *dependencies* of packages?
On Monday, August 17, 2015 at 2:34:19 PM UTC-7, Jeroen Demeyer wrote: > > On 2015-08-17 19:01, William Stein wrote: > > Let's at least establish what people are likely to think "sage -i" > > already is (after 10+ years of Sage). > I guess this is one of the questions: should Sage users be encouraged to switch from "sage -i pkg" to "make pkg" to install packages? As long as "sage -i pkg" continues to work (either as before or including dependency-checking), would "make pkg" be better for the future? I can see an argument for that. My current vote (although I keep changing my mind): - document "make pkg" to install packages, not "sage -i pkg" As far as whether "sage -i pkg" should check dependencies, I don't feel strongly about it. I don't think I want a new option "sage -m pkg" which would just run "make pkg". -- John -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Should sage -i install *dependencies* of packages?
On 2015-08-17 19:01, William Stein wrote: Let's at least establish what people are likely to think "sage -i" already is (after 10+ years of Sage). Personally, I see it as a low-level command, analogous to other low-level commands like ./sage -b and ./sage --docbuild. -i [packages] -- install the given Sage packages ... It's description is "install the given Sage packages". Wikipedia: "A package manager [...] is a collection of software tools that automates the process of installing [...] software packages..." It's not because "sage -i" installs packages and a package manager installs packages, that "sage -i" is a package manager. I'm not saying this to be pedantic, since I think both install-with-dependencies-like-a-package-manager and install-just-the-one-package are needed commands for Sage and they both install packages. Jeroen. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Should sage -i install *dependencies* of packages?
On 2015-08-17 19:12, kcrisman wrote: How many packages currently have such (additional) dependencies now as opposed to several years ago? Currently, the only packages which depend on a package which is not standard are: * python_igraph depending on igraph * qepcad depending on saclib By a strange coincidence, both are introduced in the latest beta (6.9.beta2). I only expect it to install one package at a time but I also can't remember installing a lot of packages that had dependencies that weren't standard packages. Note that "./sage -i" might also be used for standard packages when doing Sage development, it's not just optional/experimental packages. In which event the default could indeed change to installing dependencies That's obvious, the default absolutely must change to installing dependencies. But, like William's answer, that's just shifting the question: the default of *which command* should be like that? Jeroen. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Re: Should sage -i install *dependencies* of packages?
> > What would you like? > As your ticket aims at making users install packages in a way that takes dependencies into account, I vote for (A) which does not change the command that they are used to type to install packages. Nathann -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Should sage -i install *dependencies* of packages?
How many packages currently have such (additional) dependencies now as opposed to several years ago? I only expect it to install one package at a time but I also can't remember installing a lot of packages that had dependencies that weren't standard packages. Maybe we're separating out things more than we used to, so this is causing trouble for that reason? In which event the default could indeed change to installing dependencies (which I have to say is VERY convenient in R, but is also a zillion times more necessary in R). -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Should sage -i install *dependencies* of packages?
On Mon, Aug 17, 2015 at 10:48 AM, Jeroen Demeyer wrote: > On 2015-08-17 18:35, William Stein wrote: >> >> On Mon, Aug 17, 2015 at 10:00 AM, Jeroen Demeyer >> wrote: >> [...] >>> >>> What would you like? >> >> >> What does every other package manager do? > > > I don't think that's relevant. The question is not whether or not it should > be possible to install packages with dependencies (what a package manager > does), only what the user interface should be. > > My question could be rephrased as: should "sage -i" start behaving like a > package manager or should it be just a low-level command to install a single > package? Let's at least establish what people are likely to think "sage -i" already is (after 10+ years of Sage).I double people think it's just a low-level command, since "sage --help" lists it at the top level (no need to use "sage -advanced" to find out about it). sage --help ... -advanced -- list all command line options ... -i [packages] -- install the given Sage packages ... It's description is "install the given Sage packages". Wikipedia: "A package manager [...] is a collection of software tools that automates the process of installing [...] software packages..." William > > > Jeroen. > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To post to this group, send email to sage-devel@googlegroups.com. > Visit this group at http://groups.google.com/group/sage-devel. > For more options, visit https://groups.google.com/d/optout. -- William (http://wstein.org) -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Should sage -i install *dependencies* of packages?
On 2015-08-17 18:35, William Stein wrote: On Mon, Aug 17, 2015 at 10:00 AM, Jeroen Demeyer wrote: [...] What would you like? What does every other package manager do? I don't think that's relevant. The question is not whether or not it should be possible to install packages with dependencies (what a package manager does), only what the user interface should be. My question could be rephrased as: should "sage -i" start behaving like a package manager or should it be just a low-level command to install a single package? Jeroen. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Should sage -i install *dependencies* of packages?
On Mon, Aug 17, 2015 at 10:00 AM, Jeroen Demeyer wrote: [...] > What would you like? What does every other package manager do? (I am being somewhat rhetorical because I think I know the answer.) Maybe we should do what *everybody* else does in this case, to minimize surprise and confusion...? -- William -- William (http://wstein.org) -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Building Sage on Cygwin: Possibly Linking Error
I'm first trying to build with the current Cython version. If that doesn't work, then next I will try that. Best, Travis On Monday, August 17, 2015 at 9:05:34 AM UTC-7, Dima Pasechnik wrote: > > > > On Monday, 17 August 2015 08:56:27 UTC-7, Travis Scrimshaw wrote: >> >> I'm using gcc version 4.9.3.s >>> >>> Is it able to build Sage on Linux? >>> >> >> I currently don't have access to a Linux system to test it, and I >> probably won't for another 1-2 weeks. >> > > did you try building sage's gcc (4.9.2) and proceed with it? > > >> >> Best, >> Travis >> >> -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] trac chronic health problem
Fantastic. Thanks! On 17/08/15 19:18, R. Andrew Ohana wrote: There was a cron job setup for cleaning /tmp, but I just looked and it seems like it was disabled... I re-enabled it and manually cleaned it at the same time. On Sun, Aug 16, 2015 at 6:48 AM, Vincent Delecroix < 20100.delecr...@gmail.com> wrote: Hello, Again... """ Trac detected an internal error: OSError: [Errno 2] No such file or directory: '/tmp/tmpVRSiHH' """ Wouldn't be possible to set a cron job cleaning /tmp/? Vincent -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] trac chronic health problem
There was a cron job setup for cleaning /tmp, but I just looked and it seems like it was disabled... I re-enabled it and manually cleaned it at the same time. On Sun, Aug 16, 2015 at 6:48 AM, Vincent Delecroix < 20100.delecr...@gmail.com> wrote: > Hello, > > Again... > > """ > Trac detected an internal error: > > OSError: [Errno 2] No such file or directory: '/tmp/tmpVRSiHH' > """ > > Wouldn't be possible to set a cron job cleaning /tmp/? > > Vincent > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To post to this group, send email to sage-devel@googlegroups.com. > Visit this group at http://groups.google.com/group/sage-devel. > For more options, visit https://groups.google.com/d/optout. > -- Andrew -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Re: Trac Server Issue
This should now be resolved. On Sun, Aug 16, 2015 at 6:10 PM, Kwankyu Lee wrote: > I looked into this thread because I have the same issue with the following > ticket. > > http://trac.sagemath.org/ticket/18807 > > Anyway, your ticket URL seems to have no problem :-) > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To post to this group, send email to sage-devel@googlegroups.com. > Visit this group at http://groups.google.com/group/sage-devel. > For more options, visit https://groups.google.com/d/optout. > -- Andrew -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
[sage-devel] Should sage -i install *dependencies* of packages?
This question is because of discussion at #18859.
The current behaviour of "sage -i PKG" is to install package PKG (if the
latest version of that package is not already installed), without
looking at the dependencies of PKG.
Now what should the recommended user interface be for installing a
package *with* dependencies?
(A) change "sage -i PKG" to also install dependencies (adding some
option --no-dependencies to keep the old behaviour)
(B) add a new option, say "sage -m PKG" (-m for make) to install a
package with dependencies
(C) "make PKG" (which is currently supported but documented)
* Note that (C) is absolutely no change in functionality, it's just a
documentation change. (B) is a minor change in functionality (a few
lines in src/bin/sage), while (A) adds significant complexity.
* Options (B) and (C) are fully backwards compatible, no changes to
current behaviour are needed.
* An advantage of (A) and (B) is that they can be run from anywhere,
while (C) only works from $SAGE_ROOT.
* The main advantage of (A) is that users do not need to change
behaviour: usually installing packages *with* dependencies is what you want.
* The advantage of (C) is that it is consistent with "sage -b" and "sage
--docbuild FOO" which also do not install dependencies of the Sage
library or the documentation. The corresponding make targets "build" and
"doc" do build the dependencies.
* When choosing (A), we need to define what "sage -i -f PKG" and "sage
-f PKG" should do. I see the following possible variants:
(A1) Disallow "sage -i -f PKG" and "sage -f PKG" without the
--no-depencies option
(A2) The "-f" option implies --no-dependencies
(A3) The "-f" option forces re-installation of the package and all its
dependencies
(A4) Make a difference between "sage -f PKG" and "sage -i -f PKG"
My preference goes to either (B) or (C), mainly because of the added
complexity that (A) would require and because I like to keep a simple
command ("sage -i PKG") to install just package PKG.
What would you like?
Jeroen.
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Building Sage on Cygwin: Possibly Linking Error
On Monday, 17 August 2015 08:56:27 UTC-7, Travis Scrimshaw wrote: > > I'm using gcc version 4.9.3.s >>> >> >> Is it able to build Sage on Linux? >> > > I currently don't have access to a Linux system to test it, and I probably > won't for another 1-2 weeks. > did you try building sage's gcc (4.9.2) and proceed with it? > > Best, > Travis > > -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] Building Sage on Cygwin: Possibly Linking Error
> > I'm using gcc version 4.9.3.s >> > > Is it able to build Sage on Linux? > I currently don't have access to a Linux system to test it, and I probably won't for another 1-2 weeks. Best, Travis -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] matrix.subs()
On 17/08/15 09:27, David Roe wrote: On Sat, Aug 15, 2015 at 3:00 AM, Vincent Delecroix <20100.delecr...@gmail.com> wrote: Hello, As mentioned by Søren on ask [1] the subs method of matrices behave differently than the subs method on coefficients sage: R. = PolynomialRing(ZZ) sage: m = matrix(R, [[x]]) sage: x.subs(3).parent() # parent *does* change Integer Ring sage: m.subs(3).parent() # parent *does not* change Full MatrixSpace of 1 by 1 dense matrices over Univariate Polynomial Ring in x over Integer Ring Shouldn't the subs method on matrices simply be return matrix([a.subs(*args, **kwds) for a in self.list()]) instead of keeping the base ring? Sounds like a good idea to me. Ticket #19045 needs review! Vincent -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
Re: [sage-devel] matrix.subs()
On Sat, Aug 15, 2015 at 3:00 AM, Vincent Delecroix <20100.delecr...@gmail.com> wrote: > Hello, > > As mentioned by Søren on ask [1] the subs method of matrices behave > differently than the subs method on coefficients > > sage: R. = PolynomialRing(ZZ) > sage: m = matrix(R, [[x]]) > > sage: x.subs(3).parent() # parent *does* change > Integer Ring > > sage: m.subs(3).parent() # parent *does not* change > Full MatrixSpace of 1 by 1 dense matrices over > Univariate Polynomial Ring in x over Integer Ring > > Shouldn't the subs method on matrices simply be > > return matrix([a.subs(*args, **kwds) for a in self.list()]) > > instead of keeping the base ring? Sounds like a good idea to me. David > > Vincent > > [1] http://ask.sagemath.org/question/28844 > > -- > You received this message because you are subscribed to the Google Groups > "sage-devel" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to sage-devel+unsubscr...@googlegroups.com. > To post to this group, send email to sage-devel@googlegroups.com. > Visit this group at http://groups.google.com/group/sage-devel. > For more options, visit https://groups.google.com/d/optout. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To post to this group, send email to sage-devel@googlegroups.com. Visit this group at http://groups.google.com/group/sage-devel. For more options, visit https://groups.google.com/d/optout.
