search

kevinushey / longjmp

Test package for R unwind + C++ exceptions.
3 stars 0 forks source link

Not reproducable #1

Closed ltierney closed 6 years ago

ltierney commented 6 years ago

I'm not seeing the issues in the README on Linux or macOS. Can you provide a reproducible example?

The WIndows issue seems unrelated.

ltierney commented 6 years ago

I'm not seeing the issues in the README on Linux or macOS. Can you provide a reproducible example?

The WIndows issue seems unrelated.

kevinushey commented 6 years ago

Thanks for taking a look! It looks like this only reproduces when running R within RStudio. I'm not yet sure why this would make a difference.

My only guess is that perhaps RStudio is doing something behind the scenes that could be causing this. Maybe we're accidentally screwing up the protection stack, or something similar?

ltierney commented 6 years ago

Thanks -- I can reproduce this now. Apparently we allow R_ReturnedValue to be NULL (i.e. a C NULL pointer). I can work around that but want to figure out why we allow this -- there may be a good reason but it seems like a Really Bad Idea.

Best,

luke

On Thu, 1 Feb 2018, Kevin Ushey wrote:

Thanks for taking a look! It looks like this only reproduces when running R within RStudio. I'm not yet sure why this would make a difference.

My only guess is that perhaps RStudio is doing something behind the scenes that could be causing this. Maybe we're accidentally screwing up the protection stack, or something similar?

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub, or mute the thread.[ADvaVE3yQCAXezYpElDtWTGkNSP5TjwDks5tQi10gaJpZM4R2J0b.gif]

-- Luke Tierney Ralph E. Wareham Professor of Mathematical Sciences University of Iowa Phone: 319-335-3386 Department of Statistics and Fax: 319-335-3017 Actuarial Science 241 Schaeffer Hall email: luke-tierney@uiowa.edu Iowa City, IA 52242 WWW: http://www.stat.uiowa.edu

ltierney commented 6 years ago

Should be fixed in r74195. Let me know if you see any other issues.

[Remember that for this to work as intended you need the R_ContinueUnwind() call in the catch clause.]

Best,

luke

On Fri, 2 Feb 2018, luke-tierney@uiowa.edu wrote:

Thanks -- I can reproduce this now. Apparently we allow R_ReturnedValue to be NULL (i.e. a C NULL pointer). I can work around that but want to figure out why we allow this -- there may be a good reason but it seems like a Really Bad Idea.

Best,

luke

On Thu, 1 Feb 2018, Kevin Ushey wrote:

Thanks for taking a look! It looks like this only reproduces when running R within RStudio. I'm not yet sure why this would make a difference.

My only guess is that perhaps RStudio is doing something behind the scenes that could be causing this. Maybe we're accidentally screwing up the protection stack, or something similar?

— You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub, or mute the thread.[ADvaVE3yQCAXezYpElDtWTGkNSP5TjwDks5tQi10gaJpZM4R2J0b.gif]

-- Luke Tierney Ralph E. Wareham Professor of Mathematical Sciences University of Iowa Phone: 319-335-3386 Department of Statistics and Fax: 319-335-3017 Actuarial Science 241 Schaeffer Hall email: luke-tierney@uiowa.edu Iowa City, IA 52242 WWW: http://www.stat.uiowa.edu

kevinushey commented 6 years ago

Awesome news! I can confirm that all looks well with the latest version of R-devel. Thank you for taking the time to dig into this issue (and I'll fix that up in my own example).