[julia-users] Re: What to do with those Gadfly warning/error messages?

[Andreas Lobinger]

These are Glib or Gtk warnings and in the first entry you see, which 
program/process caused them. Strange is, that you get them in your julia 
REPL. I see similar warnings in the terminals, where i started e.g. firefox.

On Tuesday, April 26, 2016 at 4:06:09 AM UTC+2, K leo wrote:
>
> Plots work, but there are a lot of messages print out to REPL.  What can I 
> do about them?  See below.
>
>
> $ juliag
>_
>_   _ _(_)_ |  A fresh approach to technical computing
>   (_) | (_) (_)|  Documentation: http://docs.julialang.org
>_ _   _| |_  __ _   |  Type "?help" for help.
>   | | | | | | |/ _` |  |
>   | | |_| | | | (_| |  |  Version 0.4.5 (2016-03-18 00:58 UTC)
>  _/ |\__'_|_|_|\__'_|  |  Official http://julialang.org/ release
> |__/   |  x86_64-unknown-linux-gnu
>
> julia> using Plots
>
> julia> gadfly()
> Plots.GadflyBackend()
>
> julia> plot(x=rand(10), y=rand(10))
> [Plots.jl] Initializing backend: gadfly
> julia> 
> (Thunar:4638): GLib-GObject-WARNING **: invalid unclassed pointer in cast 
> to 'GObject'
>
> (Thunar:4638): GLib-GObject-CRITICAL **: g_object_set_qdata: assertion 
> 'G_IS_OBJECT (object)' failed
>
> (firefox:4651): Gtk-WARNING **: Theme directory  of theme oxygen has no 
> size field
>
>

[julia-users] What to do with those Gadfly warning/error messages?

[K leo]

Plots work, but there are a lot of messages print out to REPL.  What can I 
do about them?  See below.


$ juliag
   _
   _   _ _(_)_ |  A fresh approach to technical computing
  (_) | (_) (_)|  Documentation: http://docs.julialang.org
   _ _   _| |_  __ _   |  Type "?help" for help.
  | | | | | | |/ _` |  |
  | | |_| | | | (_| |  |  Version 0.4.5 (2016-03-18 00:58 UTC)
 _/ |\__'_|_|_|\__'_|  |  Official http://julialang.org/ release
|__/   |  x86_64-unknown-linux-gnu

julia> using Plots

julia> gadfly()
Plots.GadflyBackend()

julia> plot(x=rand(10), y=rand(10))
[Plots.jl] Initializing backend: gadfly
julia> 
(Thunar:4638): GLib-GObject-WARNING **: invalid unclassed pointer in cast 
to 'GObject'

(Thunar:4638): GLib-GObject-CRITICAL **: g_object_set_qdata: assertion 
'G_IS_OBJECT (object)' failed

(firefox:4651): Gtk-WARNING **: Theme directory  of theme oxygen has no 
size field

[julia-users] Improvements to GPU integration with Julia

[Michael Jin]

(Reposted from julia-dev I was told that julia-user was a 
more appropriate place to have this thread.)

Hi, I'm an undergraduate student and I've been using Julia since 2013. I've 
been trying to use the GPU seamlessly for projects involving Julia 
matrices. For that end, I have started working on my own OpenCL BLAS Julia 
library to test the clBLAS library at the lowest level possible for the GPU 
with the OpenCL C library.

Here's a link to my project: https://github.com/mikhail-j/OpenCLBLAS.jl

This project has been tested on a NVIDIA GTX 780 Ti.

Any suggestions on what I can do to improve the state of GPU integration 
with Julia?

[julia-users] Installation Problem

[Nadir]

Hi,

I tried to install Juno,when trying to add uber-juno in Atom, I get the 
following message:

Go to the Packages → Julia → Open Terminal menu and
run `Pkg.add("Atom")` in Julia, then try again.
If you still see an issue, please report it to:
julia-users@googlegroups.com

After trying to add the package manually in the shell, I get the following 
message after 30 mins ...


julia> Pkg.add("Atom")
INFO: Initializing package repository Q:\.julia\v0.4
INFO: Cloning METADATA from git://github.com/JuliaLang/METADATA.jl
fatal: The remote end hung up unexpectedly
ERROR: failed process: Process(`git clone -q -b metadata-v2 git://
github.com/Jul
iaLang/METADATA.jl METADATA`, ProcessExited(128)) [128]
 in run at process.jl:531

julia>

Is this a problem on my end/in my installation of atom?

Thanks for your help and consideration.

[julia-users] Syntax highlighting for julia symbols in Notepad++?

[Alex Mellnik]

There's a Notepad++ configuration file provided in the distribution here 
. 
 One limitation is that it doesn't highlight symbols (like *:something*) 
and always treats : as an operator.  IJulia on the other hand is smart 
enough to highlight them with a custom style.  Is there a way to do this in 
Notepad++?  I haven't been able to get it to work without catching at least 
one trailing character.  -A

Re: [julia-users] Re: Arrow Keys when debugging

[chobbes158]

Many thanks!

On Monday, April 25, 2016 at 10:39:08 PM UTC+1, Stefan Karpinski wrote:
>
> Yes, to all of the above.
>
> On Mon, Apr 25, 2016 at 5:34 PM,  wrote:
>
>> Thanks a lot for the info. 
>>
>> 1. I've been using Matlab debugger and gdb for fortran for many years. 
>> Will Gallium and AST Interpreter for Julia v5.0 give a comparable user 
>> experience? For example, I'm used to use debugger to step a code from line 
>> 1 through the end. Will Gallium and ASTI allow me to do so? Debug.jl only 
>> allows debugging in functions and it cannot be used to explore ctors, if I 
>> understand correctly. Due to the nature of multidispatch, I really want to 
>> see which constructor is called for construction, if multiple constructors 
>> exist.
>>
>> 2. Will there be any proper documentation for the new debugger? 
>>
>> Sorry for being a bit annoying. Thanks!!
>>
>>
>>
>> On Monday, April 25, 2016 at 9:26:16 PM UTC+1, Keno Fischer wrote:
>>>
>>> On Mon, Apr 25, 2016 at 11:12 AM,   wrote: 
>>> > Thanks for the reference. I quickly scanned the Readmes of Gallium and 
>>> > ASTInterpreter and found that they're probably too brief to start 
>>> with. Is 
>>> > there a more complete documentation or tutorial? What is the relation 
>>> > between Gallium and AST Interpreter? 
>>>
>>> ASTInterpreter is essentially the conceptual replacement of Debug.jl. 
>>> It takes a julia function and lets you step through it. Gallium adds 
>>> breakpointing, backtracing and other 
>>> features to make it a full debugger. You can use ASTInterpreter 
>>> without Gallium, but the 
>>> otherway around is not super useful. 
>>>
>>> > BTW, do I have to wait until the release of Julia v5.0 to use Gallium? 
>>>
>>> Gallium is available on currently julia master, but not on 0.4. 
>>>
>>> > Sorry, if my questions look dumb. 
>>> > 
>>> > Thanks!! 
>>> > 
>>> > 
>>> > 
>>> > 
>>> > 
>>> > On Monday, April 25, 2016 at 3:20:33 PM UTC+1, Steven G. Johnson 
>>> wrote: 
>>> >> 
>>> >> 
>>> >> 
>>> >> On Monday, April 25, 2016 at 10:10:01 AM UTC-4, chobb...@gmail.com 
>>> wrote: 
>>> >>> 
>>> >>> I'm using the Julia Debugger 
>>> >>> 
>>> >>> https://github.com/toivoh/Debug.jl 
>>> >> 
>>> >> 
>>> >> That debugger is mostly a stopgap/prototype until we have a "real" 
>>> >> debugger based on LLDB, which allows you to run code at full speed 
>>> until you 
>>> >> hit a break point.  Such a debugger was recently introduced for Julia 
>>> 0.5, 
>>> >> and is much more full-featured than Debug.jl: 
>>> >> 
>>> >> https://github.com/Keno/Gallium.jl 
>>>
>>
>

[julia-users] Re: Do I have to workspace() when a new breakpoint is introduced?

[Adham Beyki]

>
> > It seems to me that I have to clear everything in the workspace by 
> workspace() if I want to introduce new breakpoints 
>

If the file you are debugging doesn't belong to any of the modules that you 
use, you don't have to clear everything. Separate the part that you declare 
using the modules and the function you are debugging. This way you will 
have to load the packages once. After adding/removing the breakpoints, you 
only need to reload the function you are debugging and since it doesn't 
have any "using" it will take much less time.

Re: [julia-users] Re: Arrow Keys when debugging

[Stefan Karpinski]

Yes, to all of the above.

On Mon, Apr 25, 2016 at 5:34 PM,  wrote:

> Thanks a lot for the info.
>
> 1. I've been using Matlab debugger and gdb for fortran for many years.
> Will Gallium and AST Interpreter for Julia v5.0 give a comparable user
> experience? For example, I'm used to use debugger to step a code from line
> 1 through the end. Will Gallium and ASTI allow me to do so? Debug.jl only
> allows debugging in functions and it cannot be used to explore ctors, if I
> understand correctly. Due to the nature of multidispatch, I really want to
> see which constructor is called for construction, if multiple constructors
> exist.
>
> 2. Will there be any proper documentation for the new debugger?
>
> Sorry for being a bit annoying. Thanks!!
>
>
>
> On Monday, April 25, 2016 at 9:26:16 PM UTC+1, Keno Fischer wrote:
>>
>> On Mon, Apr 25, 2016 at 11:12 AM,   wrote:
>> > Thanks for the reference. I quickly scanned the Readmes of Gallium and
>> > ASTInterpreter and found that they're probably too brief to start with.
>> Is
>> > there a more complete documentation or tutorial? What is the relation
>> > between Gallium and AST Interpreter?
>>
>> ASTInterpreter is essentially the conceptual replacement of Debug.jl.
>> It takes a julia function and lets you step through it. Gallium adds
>> breakpointing, backtracing and other
>> features to make it a full debugger. You can use ASTInterpreter
>> without Gallium, but the
>> otherway around is not super useful.
>>
>> > BTW, do I have to wait until the release of Julia v5.0 to use Gallium?
>>
>> Gallium is available on currently julia master, but not on 0.4.
>>
>> > Sorry, if my questions look dumb.
>> >
>> > Thanks!!
>> >
>> >
>> >
>> >
>> >
>> > On Monday, April 25, 2016 at 3:20:33 PM UTC+1, Steven G. Johnson wrote:
>> >>
>> >>
>> >>
>> >> On Monday, April 25, 2016 at 10:10:01 AM UTC-4, chobb...@gmail.com
>> wrote:
>> >>>
>> >>> I'm using the Julia Debugger
>> >>>
>> >>> https://github.com/toivoh/Debug.jl
>> >>
>> >>
>> >> That debugger is mostly a stopgap/prototype until we have a "real"
>> >> debugger based on LLDB, which allows you to run code at full speed
>> until you
>> >> hit a break point.  Such a debugger was recently introduced for Julia
>> 0.5,
>> >> and is much more full-featured than Debug.jl:
>> >>
>> >> https://github.com/Keno/Gallium.jl
>>
>

Re: [julia-users] Re: Arrow Keys when debugging

[chobbes158]

Thanks a lot for the info. 

1. I've been using Matlab debugger and gdb for fortran for many years. Will 
Gallium and AST Interpreter for Julia v5.0 give a comparable user 
experience? For example, I'm used to use debugger to step a code from line 
1 through the end. Will Gallium and ASTI allow me to do so? Debug.jl only 
allows debugging in functions and it cannot be used to explore ctors, if I 
understand correctly. Due to the nature of multidispatch, I really want to 
see which constructor is called for construction, if multiple constructors 
exist.

2. Will there be any proper documentation for the new debugger? 

Sorry for being a bit annoying. Thanks!!



On Monday, April 25, 2016 at 9:26:16 PM UTC+1, Keno Fischer wrote:
>
> On Mon, Apr 25, 2016 at 11:12 AM,   
> wrote: 
> > Thanks for the reference. I quickly scanned the Readmes of Gallium and 
> > ASTInterpreter and found that they're probably too brief to start with. 
> Is 
> > there a more complete documentation or tutorial? What is the relation 
> > between Gallium and AST Interpreter? 
>
> ASTInterpreter is essentially the conceptual replacement of Debug.jl. 
> It takes a julia function and lets you step through it. Gallium adds 
> breakpointing, backtracing and other 
> features to make it a full debugger. You can use ASTInterpreter 
> without Gallium, but the 
> otherway around is not super useful. 
>
> > BTW, do I have to wait until the release of Julia v5.0 to use Gallium? 
>
> Gallium is available on currently julia master, but not on 0.4. 
>
> > Sorry, if my questions look dumb. 
> > 
> > Thanks!! 
> > 
> > 
> > 
> > 
> > 
> > On Monday, April 25, 2016 at 3:20:33 PM UTC+1, Steven G. Johnson wrote: 
> >> 
> >> 
> >> 
> >> On Monday, April 25, 2016 at 10:10:01 AM UTC-4, chobb...@gmail.com 
> wrote: 
> >>> 
> >>> I'm using the Julia Debugger 
> >>> 
> >>> https://github.com/toivoh/Debug.jl 
> >> 
> >> 
> >> That debugger is mostly a stopgap/prototype until we have a "real" 
> >> debugger based on LLDB, which allows you to run code at full speed 
> until you 
> >> hit a break point.  Such a debugger was recently introduced for Julia 
> 0.5, 
> >> and is much more full-featured than Debug.jl: 
> >> 
> >> https://github.com/Keno/Gallium.jl 
>

[julia-users] Why does QR beat LU for sparse, square matrices?

[Jonas Kersulis]

Can someone explain why qrfact is faster and requires less memory than 
lufact for square, sparse matrices? LU is less computationally complex than 
QR, and with decent pivoting LU should be able to preserve sparsity better 
than QR, so I'm confused by what I'm seeing in practice.

The plots below compare computation time, memory allocation, and garbage 
collection time for the two factorization methods. I generated them by 
factorizing sprand sparse matrices. The top plot shows results for matrices 
with 10% nonzeros; the bottom plot shows results for matrices with 50% 
nonzeros. The methods seem to converge in memory performance as density 
increases, but LU loses to QR in both cases.

I have looked through the paper describing the multifrontal QR algorithm 
Julia calls, but I don't understand it fully. Before I spend more time 
studying it, I figured I would see if someone here knows the secret sauce 
that helps it beat LU.

[julia-users] [ANN] QML.jl progress

[Bart Janssens]

Hi all,

I just wanted to report on some progress for QML.jl at 
https://github.com/barche/QML.jl :
- Communication mechanisms with Julia for basic types (signals, function 
calls and context properties)
- Display for PNG mime type (see plot.jl example using Plots.jl and the 
image test)
- Integration into Julia event loop for combining with the REPL (e.g. to 
display a plot in a part of the QML interface)

There are experimental Windows binaries, but I haven't been able to test 
them on anything except my development machine.

Next thing I want to work on is GLVisualize integration, since it is 
possible to provide an OpenGL context from QML using a 
QQuickFramebufferObject.

Cheers,

Bart

Re: [julia-users] Re: Arrow Keys when debugging

[Keno Fischer]

On Mon, Apr 25, 2016 at 11:12 AM,   wrote:
> Thanks for the reference. I quickly scanned the Readmes of Gallium and
> ASTInterpreter and found that they're probably too brief to start with. Is
> there a more complete documentation or tutorial? What is the relation
> between Gallium and AST Interpreter?

ASTInterpreter is essentially the conceptual replacement of Debug.jl.
It takes a julia function and lets you step through it. Gallium adds
breakpointing, backtracing and other
features to make it a full debugger. You can use ASTInterpreter
without Gallium, but the
otherway around is not super useful.

> BTW, do I have to wait until the release of Julia v5.0 to use Gallium?

Gallium is available on currently julia master, but not on 0.4.

> Sorry, if my questions look dumb.
>
> Thanks!!
>
>
>
>
>
> On Monday, April 25, 2016 at 3:20:33 PM UTC+1, Steven G. Johnson wrote:
>>
>>
>>
>> On Monday, April 25, 2016 at 10:10:01 AM UTC-4, chobb...@gmail.com wrote:
>>>
>>> I'm using the Julia Debugger
>>>
>>> https://github.com/toivoh/Debug.jl
>>
>>
>> That debugger is mostly a stopgap/prototype until we have a "real"
>> debugger based on LLDB, which allows you to run code at full speed until you
>> hit a break point.  Such a debugger was recently introduced for Julia 0.5,
>> and is much more full-featured than Debug.jl:
>>
>> https://github.com/Keno/Gallium.jl

RE: [julia-users] Cloning private package with 0.5

[David Anthoff]

https://github.com/JuliaLang/julia/issues/16041

 

We can continue this discussion there. I think that issue should get a 0.5.0 
label, i.e. this should be treated as a regression that has to be fixed before 
0.5 can be released.

 

Cheers,

David

 

From: julia-users@googlegroups.com [mailto:julia-users@googlegroups.com] On 
Behalf Of David Anthoff
Sent: Monday, April 25, 2016 11:16 AM
To: julia-users@googlegroups.com
Subject: RE: [julia-users] Cloning private package with 0.5

 

That doesn’t work for bitbucket repos, though…

 

From: karbar...@gmail.com   
[mailto:karbar...@gmail.com] On Behalf Of Jacob Quinn
Sent: Sunday, April 24, 2016 8:49 PM
To: julia-users@googlegroups.com  
Subject: Re: [julia-users] Cloning private package with 0.5

 

Note that it's currently possible to clone private GitHub repos using "personal 
access tokens" from Github. If you go to your profile/settings, you can 
generate private access tokens that allow access to private Github repos and 
then from Julia, you can clone with the following format:

 

Pkg.clone("https://username:to...@github.com/username/private_repo.jl;)

 

On Fri, Apr 22, 2016 at 5:57 PM, Eric Forgy  > wrote:

I have the same problem (on Windows) and really hoping this gets fixed before 
0.5 is released. I have a ton of private repos.

I'd prefer a solution that lets me keep my remotes at 
g...@github.com:EricForgy/MyPrivateRepo.jl 
 . It would be more than a 
little annoying if I have to change every one to https. Hopefully, Pkg can do 
some magic and make the changes for me if needed.

 

RE: [julia-users] Cloning private package with 0.5

[David Anthoff]

That doesn’t work for bitbucket repos, though…

 

From: karbar...@gmail.com [mailto:karbar...@gmail.com] On Behalf Of Jacob Quinn
Sent: Sunday, April 24, 2016 8:49 PM
To: julia-users@googlegroups.com
Subject: Re: [julia-users] Cloning private package with 0.5

 

Note that it's currently possible to clone private GitHub repos using "personal 
access tokens" from Github. If you go to your profile/settings, you can 
generate private access tokens that allow access to private Github repos and 
then from Julia, you can clone with the following format:

 

Pkg.clone("https://username:to...@github.com/username/private_repo.jl;)

 

On Fri, Apr 22, 2016 at 5:57 PM, Eric Forgy  > wrote:

I have the same problem (on Windows) and really hoping this gets fixed before 
0.5 is released. I have a ton of private repos.

I'd prefer a solution that lets me keep my remotes at 
g...@github.com:EricForgy/MyPrivateRepo.jl 
 . It would be more than a 
little annoying if I have to change every one to https. Hopefully, Pkg can do 
some magic and make the changes for me if needed.

 

RE: [julia-users] Cloning private package with 0.5

[David Anthoff]

None of the suggestions in [1,2] work. They are all based on command line git 
using a credential helper. I have a working credential helper for command line 
git on my system, but Pkg.clone doesn’t seem to know or use that credential 
helper at all. As far as I can tell, libgit2 is not using credential helpers in 
the first place, so that whole route probably doesn’t help.

 

About SSH, is the plan to make those changes to the makefile and include 
libssh2 in the windows binary distribution, so that one can just clone private 
repos via ssh? That would be great!

 

From: julia-users@googlegroups.com [mailto:julia-users@googlegroups.com] On 
Behalf Of Isaiah Norton
Sent: Saturday, April 23, 2016 12:56 PM
To: julia-users@googlegroups.com
Subject: Re: [julia-users] Cloning private package with 0.5

 

I would hope that this is a release blocker issue, given that private package 
repos have been supported forever. 

So right now there seems no way to use them on 0.5 from Windows...

 

https:// should be supported, although I haven't tried it on Windows. See [1,2] 
for configuration details and instructions to avoid password prompts.

 

Regarding ssh, see [3]:

 

building against libssh2 is explicitly disabled by our Makefile.

 

(this line [2])

 

Switching ssh on for windows would require Makefile rules to build and package 
libssh2 along with libgit2 and the rest of the libraries. Note that libssh2 can 
use native windows crypto functionality ("WinCNG" backend), so it should not 
impose any further requirements.

 

[1] 
https://help.github.com/articles/which-remote-url-should-i-use/#cloning-with-https-urls-recommended

[2] 
http://stackoverflow.com/questions/13800289/configure-git-clients-like-github-for-windows-to-not-ask-for-authentication/18607931#18607931

[3] https://groups.google.com/d/msg/julia-users/GEwX49wwmdo/CihXC62yCAAJ

[4]  
https://github.com/JuliaLang/julia/blob/c2042326cc5d68db01c0b2e0d12a15f48b03563a/deps/Makefile#L2094

 

 

On Fri, Apr 22, 2016 at 8:30 PM, David Anthoff  > wrote:

Yes, I would hope that this is a release blocker issue, given that private 
package repos have been supported forever. I also have a very large number of 
private repos.

So right now there seems no way to use them on 0.5 from Windows...

I also would prefer to use SSH keys for authentication. If it does have to be 
https, I guess it would be ok if things are integrated with the git credential 
storage infrastructure. What would really be sort of a disaster is if I was 
asked for a username and password every time I do an Pkg.update() or something 
like that...


> -Original Message-
> From: julia-users@googlegroups.com   
> [mailto:julia-  
> us...@googlegroups.com  ] On Behalf Of Eric 
> Forgy
> Sent: Friday, April 22, 2016 4:58 PM
> To: julia-users   >
> Subject: RE: [julia-users] Cloning private package with 0.5
>
> I have the same problem (on Windows) and really hoping this gets fixed
> before 0.5 is released. I have a ton of private repos.
>
> I'd prefer a solution that lets me keep my remotes at
> g...@github.com:EricForgy/MyPrivateRepo.jl 
>  . It would be more than a 
> little
> annoying if I have to change every one to https. Hopefully, Pkg can do some
> magic and make the changes for me if needed.

 

[julia-users] Re: Trying to format and export plot in PyPlot

[Richard Melito]

Thanks!

On Monday, April 25, 2016 at 9:50:58 AM UTC-4, Steven G. Johnson wrote:
>
> To export it as a PDF, just do savefig("foo.pdf")
>

[julia-users] Re: Trying to format and export plot in PyPlot

[Richard Melito]

Thanks, I needed that! :)

On Monday, April 25, 2016 at 8:09:09 AM UTC-4, Cedric St-Jean wrote:
>
> That's just how contourf works. You can invert the y axis with 
> `ax[:invert_yaxis]()`. FYI, this is not specific to PyPlot, I found out the 
> answer googling for "matplotlib invert y axis". PyPlot uses PyCall, 
> PyCall's documentation should be useful.
>
> On Monday, April 25, 2016 at 4:47:35 AM UTC-4, Richard Melito wrote:
>>
>> Hello!
>>
>> I am trying to use PyPlot to programmatically create a filled contour 
>> plot (contourf) based on an array of data, set the size of the plot, and 
>> then export it as a PDF slide.
>>
>> So far, all I have figured out is how to create the contour plot, but the 
>> y-axis is backwards and I can't figure out how to fix it or anything else.
>>
>> Here's my code and the result:
>>
>> Arr=zeros(3,3)
>> for i=1:3
>> for j=1:3
>> Arr[i,j] = 2*i - j
>> end
>> end
>> Arr
>>
>> 3x3 Array{Float64,2}:
>>  1.0  0.0  -1.0
>>  3.0  2.0   1.0
>>  5.0  4.0   3.0
>>
>>
>> fig = figure()
>> cont = contourf(Arr, 12)
>> colorbar(cont)
>> ax = gca()
>>
>>
>>
>>
>>
>>

Re: [julia-users] Re: Grant funding and Julia Computing?

[Christoph Ortner]

ok, thank you both for this. I'll keep it in mind for the next opportunity.
Christoph


On Monday, 25 April 2016 12:49:08 UTC+1, Viral Shah wrote:
>
> +1 to what Steve says. We do a bunch of stuff with grants at Julia 
> Computing, which fund new capabilities, or building domain specific 
> packages. If you look at our blog posts, you can see the kind of work 
> funded through such grants. In the simplest case, it will greatly help both 
> parties, if grants include a small budget for buying support from Julia 
> Computing.
>
> Please reach out to me about any such collaborations or joint grants - big 
> or small.
>
> -viral
>
> On Saturday, April 23, 2016 at 2:00:27 AM UTC+5:30, Steven G. Johnson 
> wrote:
>>
>>
>>
>> On Friday, April 22, 2016 at 4:04:29 PM UTC-4, Christoph Ortner wrote:
>>>
>>> I've run into a similar problem in the past, but this was before Julia 
>>> Computing. If `Julia Professional Edition` or `Deployment Edition` 
>>> comes with some reasonable features are not included in the free version, 
>>> then maybe it could be justified?
>>>
>>
>> No, you would justify it with "Our research goal X would be greatly 
>> facilitated by implementing features Y and/or fixing bugs Z, and it is more 
>> practical/cost-effective to hire JC as consultants to do this than to try 
>> to find in-house expertise to solve the same problems."
>>
>

[julia-users] Re: Do I have to workspace() when a new breakpoint is introduced?

[chobbes158]

Yes, this is all about Debug.jl and has nothing to do with Gallium.



On Monday, April 25, 2016 at 4:03:37 PM UTC+1, Steven G. Johnson wrote:
>
>
>
> On Monday, April 25, 2016 at 7:23:06 AM UTC-4, Hans-Peter wrote:
>>
>> I think it's a bit too early for general debugger consumption (especially 
>> for someone new to Julia). The debugger most likely will come in 
>> well-rounded form with release 0.5 which, maybe, will be ready for the 
>> JuliaCon.
>>
>
> (Note that the poster is using Debug.jl, not the "real" Gallium debugger, 
> as was clarified in another thread.)
>

[julia-users] Re: Arrow Keys when debugging

[chobbes158]

Thanks for the reference. I quickly scanned the Readmes of Gallium and 
ASTInterpreter and found that they're probably too brief to start with. Is 
there a more complete documentation or tutorial? What is the relation 
between Gallium and AST Interpreter? 

BTW, do I have to wait until the release of Julia v5.0 to use Gallium?

Sorry, if my questions look dumb.

Thanks!!




On Monday, April 25, 2016 at 3:20:33 PM UTC+1, Steven G. Johnson wrote:
>
>
>
> On Monday, April 25, 2016 at 10:10:01 AM UTC-4, chobb...@gmail.com wrote:
>>
>> I'm using the Julia Debugger
>>
>> https://github.com/toivoh/Debug.jl
>>
>
> That debugger is mostly a stopgap/prototype until we have a "real" 
> debugger based on LLDB, which allows you to run code at full speed until 
> you hit a break point.  Such a debugger was recently introduced for Julia 
> 0.5, and is much more full-featured than Debug.jl:
>
> https://github.com/Keno/Gallium.jl
>
>

[julia-users] delegatin constructors

[Igor Cerovsky]

Hello,

Does Julia supports delegating constructors similar to c++11? If yes, what 
is a syntax? Thanks.

// c++ code

class class_a {
public:
class_a() {}
class_a(string str) : m_string{ str } {}
class_a(string str, double dbl) : m_string{ str }, m_double{ dbl } {}
double m_double;
string m_string;
};

[julia-users] Re: Do I have to workspace() when a new breakpoint is introduced?

[Steven G. Johnson]

On Monday, April 25, 2016 at 7:23:06 AM UTC-4, Hans-Peter wrote:
>
> I think it's a bit too early for general debugger consumption (especially 
> for someone new to Julia). The debugger most likely will come in 
> well-rounded form with release 0.5 which, maybe, will be ready for the 
> JuliaCon.
>

(Note that the poster is using Debug.jl, not the "real" Gallium debugger, 
as was clarified in another thread.)

[julia-users] Re: bicgstab

[Richard Tsai]

Thanks. Do you know why it is not merged after such a long period?

On Monday, April 25, 2016 at 2:55:01 PM UTC+2, Kristoffer Carlsson wrote:
>
> There is one sitting in a PR here. 
> https://github.com/JuliaLang/IterativeSolvers.jl/pull/4 Might be worth 
> looking at.
>
> On Monday, April 25, 2016 at 1:44:32 PM UTC+2, Richard Tsai wrote:
>>
>> Hi, is there a Julia implementation of the biconjugate gradient 
>> stabilized algorithm?
>>
>>

[julia-users] Re: Do I have to workspace() when a new breakpoint is introduced?

[chobbes158]

Thanks for the tips. It actually doesn't work, at least not on my machine. 
I try to reload the module in which new breakpoints are added and they're, 
again, ignored. Reloading the debugger doesn't help too. 

Thanks anyway.




On Monday, April 25, 2016 at 12:23:06 PM UTC+1, Hans-Peter wrote:
>
> > have to clear everything in the workspace by workspace() [...snip...] 
> ignore all my changes for debugging
>
> (remark: I didn't yet try out the debugger)
>
> You may put your code into a temporary module such that you only need to 
> reload the code you are working on, see e.g. 
> http://docs.julialang.org/en/release-0.4/manual/workflow-tips/. I'm not 
> sure if the debugger then pickups changes. 
>
> I think it's a bit too early for general debugger consumption (especially 
> for someone new to Julia). The debugger most likely will come in 
> well-rounded form with release 0.5 which, maybe, will be ready for the 
> JuliaCon.
>
>
>
>
>

[julia-users] Re: Arrow Keys when debugging

[Steven G. Johnson]

On Monday, April 25, 2016 at 10:10:01 AM UTC-4, chobb...@gmail.com wrote:
>
> I'm using the Julia Debugger
>
> https://github.com/toivoh/Debug.jl
>

That debugger is mostly a stopgap/prototype until we have a "real" debugger 
based on LLDB, which allows you to run code at full speed until you hit a 
break point.  Such a debugger was recently introduced for Julia 0.5, and is 
much more full-featured than Debug.jl:

https://github.com/Keno/Gallium.jl

Re: [julia-users] constant values from dictionary as compile time constants

[Michael Louwrens]

I was initially using a bunch of constant globals. It worked well but it was a 
pain to store all the values compared to a dictionary. I have not tried an 
immutable yet though! Completely forgot about those as an alternative. 

After reading the value types section again as well as reading some more 
examples I see where my initial understanding was off. 

Thanks for the help though! Going to test the immutable after a bit. 

> On 25 Apr 2016, at 15:56, Stefan Karpinski  wrote:
> 
> What about setting up a bunch of const globals?
> 
>> On Sun, Apr 24, 2016 at 8:43 AM, Tom Breloff  wrote:
>> If it's always the same keys, maybe build an immutable type at the very 
>> beginning?
>> 
>> 
>>> On Sunday, April 24, 2016, Michael Louwrens 
>>>  wrote:
>>> I use a dictionary to store all the various variables I use in an 
>>> algorithm. 
>>> 
>>> So, there are many functions which look like:
>>> function foo(input::Dict)
>>>x =2*input[:bar]
>>> end
>>> 
>>> Input is created once and never changed.
>>> Would abusing Val{T} be an option here? There will only be one value for 
>>> `input[:bar] for the entire run which takes > 2 hours.
>>> 
>>> Thanks!
> 

[julia-users] Re: Arrow Keys when debugging

[chobbes158]

Thanks.

I'm using the Julia Debugger

https://github.com/toivoh/Debug.jl

in a bash terminal.

Are there any other debuggers? I thought this is the only one and the de 
facto official Julia debugger.

Thanks.


On Monday, April 25, 2016 at 2:52:17 PM UTC+1, Steven G. Johnson wrote:
>
> What debugger are you using?
>

Re: [julia-users] Re: build issue on v0.4.2

[ming . chau]

my "clients" want to validate the "build & custom" automation before 
thinking about upgrading. i have no choice :)


Le lundi 25 avril 2016 14:51:38 UTC+2, Michele Zaffalon a écrit :
>
> Would it not be better to update to 0.4.5 first?
>
> On Mon, Apr 25, 2016 at 1:57 PM,  
> wrote:
>
>> it seems that
>>
>> /home/vagrant/julia-0.4.2/base/precompile.jl
>>
>> crashed because of memory overflow (i watched htop).
>>
>> i'll try the build with more memory on my test virtual machine (500Mb -> 
>> 1 Gb).
>>
>>
>>
>> Le lundi 25 avril 2016 13:50:21 UTC+2, ming...@irt-systemx.fr a écrit :
>>>
>>> hello,
>>>
>>> there a build issue on a Vagrant virtual machine (i am actually 
>>> debugging my Ansible automation) :
>>>
>>> require.jl
>>> docs/helpdb.jl
>>> docs/basedocs.jl
>>> /home/vagrant/julia-0.4.2/base/precompile.jl
>>> Killed
>>> *** This error is usually fixed by running `make clean`. If the error 
>>> persists, try `make cleanall`. ***
>>> make[1]: *** [/home/vagrant/julia-0.4.2/usr/lib/julia/sys.o] Error 1
>>> make: *** [julia-sysimg-release] Error 2
>>> vagrant@vagrant-ubuntu-trusty-64:~/julia-0.4.2$ make
>>> LINK usr/lib/julia/sys.so
>>> /home/vagrant/julia-0.4.2/usr/lib/julia/sys.o: file not recognized: File 
>>> truncated
>>> collect2: error: ld returned 1 exit status
>>> make[1]: *** [/home/vagrant/julia-0.4.2/usr/lib/julia/sys.so] Error 1
>>>
>>>
>>> and here is the sys.o file :
>>>
>>> vagrant@vagrant-ubuntu-trusty-64:~/julia-0.4.2/usr/lib/julia$ ll
>>> total 3788
>>> drwxrwxr-x 2 vagrant vagrant4096 Apr 25 13:27 ./
>>> drwxrwxr-x 4 vagrant vagrant4096 Apr 25 13:23 ../
>>> -rw-r--r-- 1 vagrant vagrant 1966777 Apr 25 13:23 inference0.ji
>>> -rw-r--r-- 1 vagrant vagrant 1898851 Apr 25 13:24 inference.ji
>>> -rw-rw-r-- 1 vagrant vagrant   0 Apr 25 13:25 sys.o
>>>
>>>
>>> i did :
>>> make -j 1
>>> in order to be sure that it is not a parallel make issue.
>>>
>>>
>>> bests,
>>>
>>
>

Re: [julia-users] constant values from dictionary as compile time constants

[Stefan Karpinski]

What about setting up a bunch of const globals?

On Sun, Apr 24, 2016 at 8:43 AM, Tom Breloff  wrote:

> If it's always the same keys, maybe build an immutable type at the very
> beginning?
>
>
> On Sunday, April 24, 2016, Michael Louwrens <
> michael.w.louwr...@outlook.com> wrote:
>
>> I use a dictionary to store all the various variables I use in an
>> algorithm.
>>
>> So, there are many functions which look like:
>> function foo(input::Dict)
>>x =2*input[:bar]
>> end
>>
>> Input is created once and never changed.
>> Would abusing Val{T} be an option here? There will only be one value for
>> `input[:bar] for the entire run which takes > 2 hours.
>>
>> Thanks!
>>
>

[julia-users] Re: bicgstab

['Lars Ruthotto' via julia-users]

There is also one in KrylovMethods.jl 
(https://github.com/lruthotto/KrylovMethods.jl)  for a single rhs and a 
block version for multiple right hand sides.

On Monday, April 25, 2016 at 7:44:32 AM UTC-4, Richard Tsai wrote:
>
> Hi, is there a Julia implementation of the biconjugate gradient stabilized 
> algorithm?
>
>

[julia-users] Re: Arrow Keys when debugging

[Steven G. Johnson]

What debugger are you using?

[julia-users] Re: Trying to format and export plot in PyPlot

[Steven G. Johnson]

To export it as a PDF, just do savefig("foo.pdf")

[julia-users] Re: bicgstab

[Kristoffer Carlsson]

There is one sitting in a PR 
here. https://github.com/JuliaLang/IterativeSolvers.jl/pull/4 Might be 
worth looking at.

On Monday, April 25, 2016 at 1:44:32 PM UTC+2, Richard Tsai wrote:
>
> Hi, is there a Julia implementation of the biconjugate gradient stabilized 
> algorithm?
>
>

Re: [julia-users] Re: build issue on v0.4.2

[Michele Zaffalon]

Would it not be better to update to 0.4.5 first?

On Mon, Apr 25, 2016 at 1:57 PM,  wrote:

> it seems that
>
> /home/vagrant/julia-0.4.2/base/precompile.jl
>
> crashed because of memory overflow (i watched htop).
>
> i'll try the build with more memory on my test virtual machine (500Mb -> 1
> Gb).
>
>
>
> Le lundi 25 avril 2016 13:50:21 UTC+2, ming...@irt-systemx.fr a écrit :
>>
>> hello,
>>
>> there a build issue on a Vagrant virtual machine (i am actually debugging
>> my Ansible automation) :
>>
>> require.jl
>> docs/helpdb.jl
>> docs/basedocs.jl
>> /home/vagrant/julia-0.4.2/base/precompile.jl
>> Killed
>> *** This error is usually fixed by running `make clean`. If the error
>> persists, try `make cleanall`. ***
>> make[1]: *** [/home/vagrant/julia-0.4.2/usr/lib/julia/sys.o] Error 1
>> make: *** [julia-sysimg-release] Error 2
>> vagrant@vagrant-ubuntu-trusty-64:~/julia-0.4.2$ make
>> LINK usr/lib/julia/sys.so
>> /home/vagrant/julia-0.4.2/usr/lib/julia/sys.o: file not recognized: File
>> truncated
>> collect2: error: ld returned 1 exit status
>> make[1]: *** [/home/vagrant/julia-0.4.2/usr/lib/julia/sys.so] Error 1
>>
>>
>> and here is the sys.o file :
>>
>> vagrant@vagrant-ubuntu-trusty-64:~/julia-0.4.2/usr/lib/julia$ ll
>> total 3788
>> drwxrwxr-x 2 vagrant vagrant4096 Apr 25 13:27 ./
>> drwxrwxr-x 4 vagrant vagrant4096 Apr 25 13:23 ../
>> -rw-r--r-- 1 vagrant vagrant 1966777 Apr 25 13:23 inference0.ji
>> -rw-r--r-- 1 vagrant vagrant 1898851 Apr 25 13:24 inference.ji
>> -rw-rw-r-- 1 vagrant vagrant   0 Apr 25 13:25 sys.o
>>
>>
>> i did :
>> make -j 1
>> in order to be sure that it is not a parallel make issue.
>>
>>
>> bests,
>>
>

[julia-users] Re: Trying to format and export plot in PyPlot

[Cedric St-Jean]

That's just how contourf works. You can invert the y axis with 
`ax[:invert_yaxis]()`. FYI, this is not specific to PyPlot, I found out the 
answer googling for "matplotlib invert y axis". PyPlot uses PyCall, 
PyCall's documentation should be useful.

On Monday, April 25, 2016 at 4:47:35 AM UTC-4, Richard Melito wrote:
>
> Hello!
>
> I am trying to use PyPlot to programmatically create a filled contour plot 
> (contourf) based on an array of data, set the size of the plot, and then 
> export it as a PDF slide.
>
> So far, all I have figured out is how to create the contour plot, but the 
> y-axis is backwards and I can't figure out how to fix it or anything else.
>
> Here's my code and the result:
>
> Arr=zeros(3,3)
> for i=1:3
> for j=1:3
> Arr[i,j] = 2*i - j
> end
> end
> Arr
>
> 3x3 Array{Float64,2}:
>  1.0  0.0  -1.0
>  3.0  2.0   1.0
>  5.0  4.0   3.0
>
>
> fig = figure()
> cont = contourf(Arr, 12)
> colorbar(cont)
> ax = gca()
>
>
>
>
>
>

[julia-users] Re: build issue on v0.4.2

[ming . chau]

it seems that

/home/vagrant/julia-0.4.2/base/precompile.jl

crashed because of memory overflow (i watched htop).

i'll try the build with more memory on my test virtual machine (500Mb -> 1 
Gb).


Le lundi 25 avril 2016 13:50:21 UTC+2, ming...@irt-systemx.fr a écrit :
>
> hello,
>
> there a build issue on a Vagrant virtual machine (i am actually debugging 
> my Ansible automation) :
>
> require.jl
> docs/helpdb.jl
> docs/basedocs.jl
> /home/vagrant/julia-0.4.2/base/precompile.jl
> Killed
> *** This error is usually fixed by running `make clean`. If the error 
> persists, try `make cleanall`. ***
> make[1]: *** [/home/vagrant/julia-0.4.2/usr/lib/julia/sys.o] Error 1
> make: *** [julia-sysimg-release] Error 2
> vagrant@vagrant-ubuntu-trusty-64:~/julia-0.4.2$ make
> LINK usr/lib/julia/sys.so
> /home/vagrant/julia-0.4.2/usr/lib/julia/sys.o: file not recognized: File 
> truncated
> collect2: error: ld returned 1 exit status
> make[1]: *** [/home/vagrant/julia-0.4.2/usr/lib/julia/sys.so] Error 1
>
>
> and here is the sys.o file :
>
> vagrant@vagrant-ubuntu-trusty-64:~/julia-0.4.2/usr/lib/julia$ ll
> total 3788
> drwxrwxr-x 2 vagrant vagrant4096 Apr 25 13:27 ./
> drwxrwxr-x 4 vagrant vagrant4096 Apr 25 13:23 ../
> -rw-r--r-- 1 vagrant vagrant 1966777 Apr 25 13:23 inference0.ji
> -rw-r--r-- 1 vagrant vagrant 1898851 Apr 25 13:24 inference.ji
> -rw-rw-r-- 1 vagrant vagrant   0 Apr 25 13:25 sys.o
>
>
> i did :
> make -j 1
> in order to be sure that it is not a parallel make issue.
>
>
> bests,
>

[julia-users] build issue on v0.4.2

[ming . chau]

hello,

there a build issue on a Vagrant virtual machine (i am actually debugging 
my Ansible automation) :

require.jl
docs/helpdb.jl
docs/basedocs.jl
/home/vagrant/julia-0.4.2/base/precompile.jl
Killed
*** This error is usually fixed by running `make clean`. If the error 
persists, try `make cleanall`. ***
make[1]: *** [/home/vagrant/julia-0.4.2/usr/lib/julia/sys.o] Error 1
make: *** [julia-sysimg-release] Error 2
vagrant@vagrant-ubuntu-trusty-64:~/julia-0.4.2$ make
LINK usr/lib/julia/sys.so
/home/vagrant/julia-0.4.2/usr/lib/julia/sys.o: file not recognized: File 
truncated
collect2: error: ld returned 1 exit status
make[1]: *** [/home/vagrant/julia-0.4.2/usr/lib/julia/sys.so] Error 1


and here is the sys.o file :

vagrant@vagrant-ubuntu-trusty-64:~/julia-0.4.2/usr/lib/julia$ ll
total 3788
drwxrwxr-x 2 vagrant vagrant4096 Apr 25 13:27 ./
drwxrwxr-x 4 vagrant vagrant4096 Apr 25 13:23 ../
-rw-r--r-- 1 vagrant vagrant 1966777 Apr 25 13:23 inference0.ji
-rw-r--r-- 1 vagrant vagrant 1898851 Apr 25 13:24 inference.ji
-rw-rw-r-- 1 vagrant vagrant   0 Apr 25 13:25 sys.o


i did :
make -j 1
in order to be sure that it is not a parallel make issue.


bests,

Re: [julia-users] Re: Grant funding and Julia Computing?

[Viral Shah]

+1 to what Steve says. We do a bunch of stuff with grants at Julia 
Computing, which fund new capabilities, or building domain specific 
packages. If you look at our blog posts, you can see the kind of work 
funded through such grants. In the simplest case, it will greatly help both 
parties, if grants include a small budget for buying support from Julia 
Computing.

Please reach out to me about any such collaborations or joint grants - big 
or small.

-viral

On Saturday, April 23, 2016 at 2:00:27 AM UTC+5:30, Steven G. Johnson wrote:
>
>
>
> On Friday, April 22, 2016 at 4:04:29 PM UTC-4, Christoph Ortner wrote:
>>
>> I've run into a similar problem in the past, but this was before Julia 
>> Computing. If `Julia Professional Edition` or `Deployment Edition` comes 
>> with some reasonable features are not included in the free version, then 
>> maybe it could be justified?
>>
>
> No, you would justify it with "Our research goal X would be greatly 
> facilitated by implementing features Y and/or fixing bugs Z, and it is more 
> practical/cost-effective to hire JC as consultants to do this than to try 
> to find in-house expertise to solve the same problems."
>

[julia-users] bicgstab

[Richard Tsai]

Hi, is there a Julia implementation of the biconjugate gradient stabilized 
algorithm?

[julia-users] Re: Do I have to workspace() when a new breakpoint is introduced?

[Hans-Peter]

> have to clear everything in the workspace by workspace() [...snip...] 
ignore all my changes for debugging

(remark: I didn't yet try out the debugger)

You may put your code into a temporary module such that you only need to 
reload the code you are working on, see e.g. 
http://docs.julialang.org/en/release-0.4/manual/workflow-tips/. I'm not 
sure if the debugger then pickups changes. 

I think it's a bit too early for general debugger consumption (especially 
for someone new to Julia). The debugger most likely will come in 
well-rounded form with release 0.5 which, maybe, will be ready for the 
JuliaCon.

[julia-users] Do I have to workspace() when a new breakpoint is introduced?

[chobbes158]

It seems to me that I have to clear everything in the workspace by 
workspace() if I want to introduce new breakpoints or remove existing 
breakpoints when debugging. Otherwise, the debugger simply ignore all my 
changes for debugging. But it takes huge amount of time when a lot of 
packages need to be loaded via "using". I wonder if there is a more 
efficient way to deal with this.

Thanks!!

[julia-users] Arrow Keys when debugging

[chobbes158]

I'm new to Julia. So bear me if the question is naive.

I'm wont to use up and down arrow keys to echo my previous commands and 
left and right keys to move in and out of a pair of parentheses I 
intentionally close first for adding contains. I found I can't do these 
when using Julia Debugger. 

Could anyone let me know if there is a way to retrieve these basic 
functionality of the arrow keys?

Thanks very much!

[julia-users] Trying to format and export plot in PyPlot

[Richard Melito]

Hello!

I am trying to use PyPlot to programmatically create a filled contour plot 
(contourf) based on an array of data, set the size of the plot, and then 
export it as a PDF slide.

So far, all I have figured out is how to create the contour plot, but the 
y-axis is backwards and I can't figure out how to fix it or anything else.

Here's my code and the result:

Arr=zeros(3,3)
for i=1:3
for j=1:3
Arr[i,j] = 2*i - j
end
end
Arr

3x3 Array{Float64,2}:
 1.0  0.0  -1.0
 3.0  2.0   1.0
 5.0  4.0   3.0


fig = figure()
cont = contourf(Arr, 12)
colorbar(cont)
ax = gca()