[R] segfault from systemfonts::system_fonts

[Dennis Fisher]

Duncan

Thanks for responding -- but your response did not help my mood.
Executing:
	.Call("_systemfonts_system_fonts_c")
triggered the segfault (as you proposed).

Dennis

Dennis Fisher MD
P < (The "P Less Than" Company)
Phone / Fax: 1-866-PLessThan (1-866-753-7784)
www.PLessThan.com




> On Nov 3, 2020, at 1:00 PM, Duncan Murdoch <murdoch.duncan using gmail.com> wrote:
> 
> On 03/11/2020 3:43 p.m., Dennis Fisher wrote:
>> R 4.0.3
>> OS X 10.15.7
>> Colleagues
>> When I run flextable, it generates a segfault.  I traced the problem to systemfonts::system_fonts()
>>>> require("systemfonts")
>>> Loading required package: systemfonts
>>>>  system_fonts()
>>> 
>>>  *** caught segfault ***
>>> address 0x0, cause 'memory not mapped'
>>> 
>>> Traceback:
>>>  1: system_fonts_c()
>>>  2: system_fonts()
>>> 
>>> Possible actions:
>>> 1: abort (with core dump, if enabled)
>>> 2: normal R exit
>>> 3: exit R without saving workspace
>>> 4: exit R saving workspace
>>> 
>> I updated my OS and R, deleted and reinstalled the systemfonts package -- problem persists.
>> I also opened Apple's font application and "validated" all font files (I have never installed any special fonts nor is there anything non-standard (e.g., Homebrew) on my system.
>> Of note, I can run other functions in systemfonts without problems -- only system_fonts triggers the segfault.
>> Another similar setup on OS X does not trigger the same problem, so the problem is more likely something in my system rather than a problem in R.
>> Does anyone have any ideas on how one might address this?
> 
> For what it's worth, I have the same R version and macOS version, and it works fine.  Debugging it will be hard:  all the work happens in a function called using
> 
>    .Call("_systemfonts_system_fonts_c")
> 
> If I could reproduce the bug and wanted to track it down, I think I'd do it by adding a bunch of Rprintf() commands into the source of _systemfonts_system_fonts_c and rebuilding the package.  It would be really tedious; I'm glad I'm not doing this!
> 
> Duncan Murdoch


	[[alternative HTML version deleted]]