[R-SIG-Mac] R.app crash report

Simon Urbanek simon.urbanek at r-project.org
Thu Jan 12 23:55:15 CET 2006


Paul,

On Jan 12, 2006, at 5:24 PM, Paul Roebuck wrote:

> On Thu, 12 Jan 2006, Simon Urbanek wrote:
>
>> On Jan 12, 2006, at 4:30 AM, David Jarvis wrote:
>>
>>> Installed R.app on powerbook 12" G4
>>>
>>> Crashed on launching
>>>
>>> Console.log reads:
>>> Mac OS X Version 10.3.5 (Build 7M34)
>>
>> Your OS is outdated, you need 10.3.9, so please update your system.
>> (BTW: This was asked and answered on this list this week already!!)
>
> Actually, this IS a bug in R.app, since if 10.3.9 is now
> required, that value should have been put in the Info.plist.
> The user would then be informed of the OS requirement rather
> than deal with the application crashing.
>
> Info.plist:
>         <key>LSMinimumSystemVersion</key>
> -        <string>10.3.0</string>
> +        <string>10.3.9</string>

It's not strictly a bug in R.app, because R.app doesn't require  
10.3.9, but the CRAN binary of R does. You can still safely compile  
your own R and use it with the existing R.app. Moreover the  
LSMinimumSystemVersion entry is rather a safeguard than a  
requirement, really. I agree that it would make sense to add a check  
in the CRAN binary installer, but the problem with that is that there  
is no simple way to tell the user what requirement is not met (that  
became possible only in 10.4 so it doesn't help, either) - it will  
simply not allow you to select any volume.

Incidentally, the incompatibility is not really intended directly, it  
comes from the fact that the CRAN binary is compiled on a 10.3.9  
system whose libraries are apparently not compatible with earlier  
versions (and this was the case even with earlier releases of R, so  
it didn't change *now* - it was only noticed recently). Still, I  
don't see a compelling reason not to update one's system, since R  
binary not running is the least drawback there. However, based on  
these requests I will update the documentation correspondingly. (It  
still startles me that exactly the same question with exactly the  
same subject and exactly the same cause is asked twice on the same  
list within 24h although answered promptly.)

Cheers,
Simon



More information about the R-SIG-Mac mailing list