[R-SIG-Mac] Issue corrected on 3.1.1 came back?

ALPEROVYCH Yan ALPEROVYCH at em-lyon.com
Mon Apr 24 14:46:25 CEST 2017 

Hi all,
The slow GUI came back with version 3.4 both official and patched. It is not present in 3.3.3.
Yan

Begin forwarded message:

From: Simon Urbanek <simon.urbanek at r-project.org<mailto:simon.urbanek at r-project.org>>
Subject: Re: [R-SIG-Mac] R 3.1.2 & Yosemite GUI slowness
Date: March 13, 2015 at 3:10:37 PM GMT+1
To: ALPEROVYCH Yan <ALPEROVYCH at em-lyon.com<mailto:ALPEROVYCH at em-lyon.com>>
Cc: David Dobolyi <dd2es at virginia.edu<mailto:dd2es at virginia.edu>>, "r-sig-mac at stat.math.ethz.ch<mailto:r-sig-mac at stat.math.ethz.ch>" <r-sig-mac at stat.math.ethz.ch<mailto:r-sig-mac at stat.math.ethz.ch>>

I have more evidence that points at Xcode. If you use the R.app GUI that was compiled at the time of R 3.1.1 - available here:

http://r.research.att.com/R-GUI-6783-3.1-mavericks-Release.dmg

then all is well. As we discussed earlier, the code did not change at all and the underlying R version is irrelevant, so the only piece that changed since July last year when this was built was the Xcode version (and thus the compilers used). It's not clear to me at all how the Xcode version can affect NSTextView so badly, so this will require some more digging. In the meantime, please use the above R.app binary.

Cheers,
Simon



On Mar 11, 2015, at 6:39 PM, ALPEROVYCH Yan <ALPEROVYCH at em-lyon.com> wrote:

Yes Simon, the 3.1.1 version works perfectly on Yosemite. This behavior started from 3.1.2 and now is through in 3.1.3.
Cheers,
Yan

On Mar 11, 2015, at 11:07 PM, David Dobolyi <dd2es at virginia.edu<mailto:dd2es at virginia.edu>> wrote:

I can confirm the problem does not occur in 3.1.1 and only started with 3.1.2. The R.app GUI version in 3.1.1 was 6784 and in 3.1.2 was 6833, so it seems like some regression was possible between builds within GUI version 1.65? Build 6912 in 3.1.3 for Mavericks still exhibits the behavior as Yan mentioned.

Best,

David

On Wed, Mar 11, 2015 at 5:58 PM, Simon Urbanek <simon.urbanek at r-project.org<mailto:simon.urbanek at r-project.org>> wrote:

On Mar 11, 2015, at 5:17 PM, ALPEROVYCH Yan <ALPEROVYCH at em-lyon.com<mailto:ALPEROVYCH at em-lyon.com>> wrote:

Hi Simon,

It may be a silly question but shouldn’t this Yosemite bug affect the the 3.1.1 version as well?


Yes it should, since the code didn't change since 3.1.1.


I tried also to run R from the terminal and found no rendering issues there,

That has nothing to do with it - R is not using any Cocoa for text on the command line - and Terminal is not likely to use NSTextView.


which points on the GUI - Yosemite related bug as you say. The R 3.1.1 version that I have (and in which this rendering problem is not present) runs with R.app GUI 1.65 (6784) x86_64-apple-darwin13.1.0. Does it make sense / is it possible to run R 3.1.3 with this particular version of GUI?


Are you saying the 3.1.1 binary didn't exhibit the behavior? That would be odd - I'll check tomorrow when I'm back at a Yosemite machine.

Cheers,
Simon



Cheers,
Yan

On Mar 11, 2015, at 9:21 PM, Simon Urbanek <simon.urbanek at r-project.org<mailto:simon.urbanek at r-project.org>> wrote:

On Mar 11, 2015, at 2:11 PM, David Dobolyi <dd2es at virginia.edu<mailto:dd2es at virginia.edu>> wrote:

Hello Yan,

I was curious if the new R update would fix this as well, so thanks for clarifying that it sadly does not.


There were no code changes on our end.

AFAICT this seems to be an issue in Apple's NSTextView Cocoa implementation in Yosemite. Pretty much all the time is spent in [NSTextView drawRect:] so Apple must have added some rather bad inefficiency to drawing text area in Yosemite. As you fill the console, the response time goes down as the text filed takes longer and longer to render. You can also see it in the editor if you add many lines. When you clear the console, it's responsible again since you don't have that many lines in the text view.

It seems like a serious bug in Yosemite. We don't control the rendering - it's all in Apple's code, so I don't see immediately how we can work around that bug.

Cheers,
Simon



Best,

David

PS. Simon, regardless of this bug, all of your hard work is very much appreciated on an invaluable program!

On Wed, Mar 11, 2015 at 5:44 AM, ALPEROVYCH Yan <ALPEROVYCH at em-lyon.com<mailto:ALPEROVYCH at em-lyon.com>> wrote:
Hi Simon & David,

Just for the record, I've tried the new 3.1.3 version (on the same Yosemite machine) and the issue is still there.

Best,

Yan Alperovych, Ph.D.
Associate Professor of Finance
EMLYON Business School
Tel.: +33 4 72 18 29 11<tel:%2B33%204%2072%2018%2029%2011>
E-mail: alperovych at em-lyon.com<mailto:alperovych at em-lyon.com><mailto:alperovych at em-lyon.com<mailto:alperovych at em-lyon.com>>
LinkedIN: fr.linkedin.com/in/yanalperovych/<http://fr.linkedin.com/in/yanalperovych/><http://fr.linkedin.com/in/yanalperovych/>

On Feb 24, 2015, at 7:28 PM, Simon Urbanek <simon.urbanek at r-project.org<mailto:simon.urbanek at r-project.org><mailto:simon.urbanek at r-project.org<mailto:simon.urbanek at r-project.org>>> wrote:

Thank you both, I'll look into this as soon as I get back from travels.
Simon


On Feb 23, 2015, at 4:23 PM, David Dobolyi <dd2es at virginia.edu<mailto:dd2es at virginia.edu><mailto:dd2es at virginia.edu<mailto:dd2es at virginia.edu>>> wrote:

Dear Simon,

Thank you for your reply. I am using the Mavericks build (running on latest Yosemite), and have noticed this issue ever since I upgraded to 3.1.2:

<Screen Shot 2015-02-23 at 4.12.35 PM.png>

As an aside, I occasionally see errors in the GUI like this, and they seem to come from other apps (e.g., Dropbox):

<Screen Shot 2015-02-23 at 4.13.09 PM.png>

More importantly, I wrote Yan about a workaround, which involves changing the max.print option from the default and limiting it to 1000. This reliably stops R from slowing down over time (although how long, I don't know). However, I can replicate the basic issue easily in R with the following code:

options(max.print=99999)
temp <- data.frame(1:100000, 1:100000)
temp

This should eventually turn up on screen, and after try typing anything in the console. Typing will be extremely delayed (e.g., type hello and it will take several seconds for each letter to show up, one at a time).

It's even possible to crash the GUI if you print something in a function called in a bootstrap, and then invoke a progress bar (e.g., the pROC package). Progress will be steady and then suddenly R will lock up. The moment you remove the print statement (which isn't actually printing since the progress bar is present), the same call works perfectly.

I'd love to have a real fix for this other than reducing max.print. I'd dig into the source myself but can't commit the time at the moment.

Best,

David



On Mon, Feb 23, 2015 at 4:00 PM, Simon Urbanek <simon.urbanek at r-project.org<mailto:simon.urbanek at r-project.org><mailto:simon.urbanek at r-project.org<mailto:simon.urbanek at r-project.org>>> wrote:
Are you using the Mavericks build of R? The SL build will do that, because SL has no AppNap support - but then you can also disable AppNap (see the corresponding threads here).

Cheers,
S


On Feb 14, 2015, at 12:43 AM, David Dobolyi <dd2es at virginia.edu<mailto:dd2es at virginia.edu><mailto:dd2es at virginia.edu<mailto:dd2es at virginia.edu>>> wrote:

Dear Yan,

Are you still seeing this issue? It has been bothering me for months now,
and I even found an instance where using progress_bar with underlying
output to the GUI will cause so much slowdown as to crash R given
sufficient time (i.e., as hidden printed output piles
up in the background, the whole process slows to a crawl).

Best,

David

_______________________________________________
R-SIG-Mac mailing list
R-SIG-Mac at r-project.org<mailto:R-SIG-Mac at r-project.org><mailto:R-SIG-Mac at r-project.org<mailto:R-SIG-Mac at r-project.org>>
https://stat.ethz.ch/mailman/listinfo/r-sig-mac





--
David Dobolyi, M.A.
PhD Candidate
Cognitive Psychology
University of Virginia
434-535-6073<tel:434-535-6073>
dd2es at virginia.edu<mailto:dd2es at virginia.edu><mailto:dd2es at virginia.edu<mailto:dd2es at virginia.edu>>

_______________________________________________
R-SIG-Mac mailing list
R-SIG-Mac at r-project.org<mailto:R-SIG-Mac at r-project.org><mailto:R-SIG-Mac at r-project.org<mailto:R-SIG-Mac at r-project.org>>
https://stat.ethz.ch/mailman/listinfo/r-sig-mac



----
Ce message electronique et tous les fichiers attaches qu'il contient sont confidentiels et destines exclusivement à l'usage de la personne à laquelle ils sont adresses. Si vous avez reçu ce message par erreur, merci de le retourner à son metteur. Les idees et opinions presentees dans ce message sont celles de son auteur, et ne representent pas necessairement celles de l'institution ou entite affiliee dont l'auteur est l'employe. La publication, l'usage, la distribution, l'impression ou la copie non autorisee de ce message et des attachements qu'il contient sont strictement interdits.

This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please return it to the sender. The ideas and views expressed in this email are solely those of its author, and do not necessarily represent the views of the institution or company of which the author is an employee. Unauthorized publication, use, distribution, printing or copying of this e-mail or any attached files is strictly forbidden.




--
David Dobolyi, M.A.
PhD Candidate
Cognitive Psychology
University of Virginia
434-535-6073<tel:434-535-6073>
dd2es at virginia.edu<mailto:dd2es at virginia.edu>



----
Ce message electronique et tous les fichiers attaches qu'il contient sont confidentiels et destines exclusivement à l'usage de la personne à laquelle ils sont adresses. Si vous avez reçu ce message par erreur, merci de le retourner à son metteur. Les idees et opinions presentees dans ce message sont celles de son auteur, et ne representent pas necessairement celles de l'institution ou entite affiliee dont l'auteur est l'employe. La publication, l'usage, la distribution, l'impression ou la copie non autorisee de ce message et des attachements qu'il contient sont strictement interdits.

This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please return it to the sender. The ideas and views expressed in this email are solely those of its author, and do not necessarily represent the views of the institution or company of which the author is an employee. Unauthorized publication, use, distribution, printing or copying of this e-mail or any attached files is strictly forbidden.




--
David Dobolyi, M.A.
PhD Candidate
Cognitive Psychology
University of Virginia
434-535-6073
dd2es at virginia.edu<mailto:dd2es at virginia.edu>


----
Ce message electronique et tous les fichiers attaches qu'il contient sont confidentiels et destines exclusivement à l'usage de la personne à laquelle ils sont adresses. Si vous avez reçu ce message par erreur, merci de le retourner à son metteur. Les idees et opinions presentees dans ce message sont celles de son auteur, et ne representent pas necessairement celles de l'institution ou entite affiliee dont l'auteur est l'employe. La publication, l'usage, la distribution, l'impression ou la copie non autorisee de ce message et des attachements qu'il contient sont strictement interdits.

This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please return it to the sender. The ideas and views expressed in this email are solely those of its author, and do not necessarily represent the views of the institution or company of which the author is an employee. Unauthorized publication, use, distribution, printing or copying of this e-mail or any attached files is strictly forbidden.


----
Ce message electronique et tous les fichiers attaches qu'il contient sont confidentiels et destines exclusivement à l'usage de la personne à laquelle ils sont adresses. Si vous avez reçu ce message par erreur, merci de le retourner à son metteur. Les idees et opinions presentees dans ce message sont celles de son auteur, et ne representent pas necessairement celles de l'institution ou entite affiliee dont l'auteur est l'employe. La publication, l'usage, la distribution, l'impression ou la copie non autorisee de ce message et des attachements qu'il contient sont strictement interdits.

This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please return it to the sender. The ideas and views expressed in this email are solely those of its author, and do not necessarily represent the views of the institution or company of which the author is an employee. Unauthorized publication, use, distribution, printing or copying of this e-mail or any attached files is strictly forbidden.

More information about the R-SIG-Mac mailing list