Re: [matplotlib-devel] Multipage pdf files

2009-01-06 Thread Jouni K . Seppänen
Jouni K. Seppänen writes: > Or perhaps the user-visible object doesn't need to be the same PdfFile > that is used internally I ended up doing this, calling the object PdfPages. I also gave it a savefig method so you can do pdfpages.savefig(figure) in addition to figure.savefig(pdfpages, format='

Re: [matplotlib-devel] John: Thoughts on a standard test system

2009-01-06 Thread John Hunter
On Mon, Dec 22, 2008 at 11:45 AM, Drain, Theodore R wrote: > John, > Sometime in January, we are going to spend some time fixing a few minor MPL > bugs we've hit and a probably work on a few enhancements (I'll send you a > list in Jan before we start anything - it's nothing major). We're also g

Re: [matplotlib-devel] John: Thoughts on a standard test system

2009-01-06 Thread Darren Dale
On Tue, Jan 6, 2009 at 10:02 AM, John Hunter wrote: > On Mon, Dec 22, 2008 at 11:45 AM, Drain, Theodore R > wrote: > > John, > > Sometime in January, we are going to spend some time fixing a few minor > MPL bugs we've hit and a probably work on a few enhancements (I'll send you > a list in Jan b

Re: [matplotlib-devel] John: Thoughts on a standard test system

2009-01-06 Thread Drain, Theodore R
OK - nose it is. How do you want to handle the dependency? My opinion is that since tests are development tools, it's not unreasonable to require that nose be installed by the developer and not as an embedded dependency in MPL (or at least that should be an option). > -Original Message---

Re: [matplotlib-devel] John: Thoughts on a standard test system

2009-01-06 Thread John Hunter
On Tue, Jan 6, 2009 at 9:58 AM, Drain, Theodore R wrote: > OK - nose it is. How do you want to handle the dependency? My opinion is > that since tests are development tools, it's not unreasonable to require that > nose be installed by the developer and not as an embedded dependency in MPL > (

[matplotlib-devel] contour plots & linestyles

2009-01-06 Thread Manuel Metz
Hi, I just noted a strange behavior of contour(). When a contour plot is created with negative values and using a single color instead of a cmap, contour _always_ uses the contour.negative_linestyle, even if linestyles are specifically provided: x = linspace(-pi,pi,100) X,Y = meshgrid(x,x)

Re: [matplotlib-devel] contour plots & linestyles

2009-01-06 Thread Eric Firing
Manuel Metz wrote: > Hi, > I just noted a strange behavior of contour(). When a contour plot is > created with negative values and using a single color instead of a cmap, > contour _always_ uses the contour.negative_linestyle, even if linestyles > are specifically provided: > > x = linspace(-p

Re: [matplotlib-devel] contour plots & linestyles

2009-01-06 Thread Manuel Metz
Eric Firing wrote: > Manuel Metz wrote: >> Hi, >> I just noted a strange behavior of contour(). When a contour plot is >> created with negative values and using a single color instead of a cmap, >> contour _always_ uses the contour.negative_linestyle, even if linestyles >> are specifically provid

Re: [matplotlib-devel] git questions

2009-01-06 Thread Michael Droettboom
I have successfully used the git mirror to commit changes to the maintenance branch. I've updated the matplotlib developer docs to describe how to do it (not that bad really), though it takes a while given the v0_98_4 "oops" branch ;) I have yet to figure out all the loop-de-loops required to

Re: [matplotlib-devel] git questions

2009-01-06 Thread Andrew Straw
Hi Mike, This sounds like good news. I am swamped right now, but I hope to get the git mirror on github up-to-date early next week when work is a little less busy. John (or any developer with SF admin capabilities), if we could set up some kind of auto-notification on SVN commits, I could possibly

Re: [matplotlib-devel] git questions

2009-01-06 Thread John Hunter
On Tue, Jan 6, 2009 at 3:07 PM, Andrew Straw wrote: > Hi Mike, This sounds like good news. I am swamped right now, but I hope > to get the git mirror on github up-to-date early next week when work is > a little less busy. > > John (or any developer with SF admin capabilities), if we could set up >