[R-SIG-Mac] upgraded to 4.5 (mac x86), errors
Jeff Newmiller
jdnewm|| @end|ng |rom dcn@d@v|@@c@@u@
Sat May 31 17:34:10 CEST 2025
Carl: "Doctor, it hurts when I don't upgrade properly..."
Duncan: "... then don't do that."
Carl: "But I didn't... rather I did"
Carl, copying ANY previously-installed packages over to the upgraded library is a no-no. Just because you did it previously and got lucky does not mean it is okay to do it whenever you upgrade... and your troubles here stem directly from this bad practice.
Put your broken library directory in the trash and start over from scratch.
On May 31, 2025 4:48:23 AM PDT, Carl Witthoft <carl using witthoft.com> wrote:
>Good guess, but that isn't what happened here. I carefully did NOT replace the "base" package set that's installed along with 4.5 . It was only the "contributed" collection that I copied over from 4.4 and then ran 'update' on. I apologize for not making that clear earlier.
>
>So, in fact the menu update command did not successfully "clean out" references to dylib files in the 4.4 directory, as the error messages showed. This despite my getting a full set of messages reporting successful download of the latest versions of all the "contributed" packages.
>
>On 5/31/25 7:34 AM, Duncan Murdoch wrote:
>> You shot yourself in the foot as Peter said.
>>
>> There are 3 classes of packages in R: base, recommended, and contributed. Base packages are really part of R itself, and are only changed when you update R.
>>
>> Contributed packages are add-ons, and can be updated independently of R. They do need to be reinstalled when an R update changes x or y in the x.y.z R version number. The update.packages() function or the corresponding menu items will do that.
>>
>> Recommended packages are contributed packages that are distributed with R. update.packages() will handle them, too.
>>
>> The base packages will never be updated by update.packages() (which is used by "update all"), so you were trying to use R 4.4 base packages with R 4.5.0.
>>
>> It would probably make sense to store base packages separately from contributed packages to avoid even the possibility of this kind of user error.
>>
>> You might think having update.packages() reinstall base packages would be a good idea, but if the base packages are messed up, R is broken, and update.packages() probably won't work properly.
>>
>> Duncan Murdoch
>>
>> On 2025-05-31 6:38 a.m., Carl Witthoft wrote:
>>> I did this sequence: copy over all 4.4 packages and then selected
>>> "update all" from the R-gui menu. Since several CRAN packages (per
>>> initial posting) still failed, I guess there's something inside
>>> update.packages that doesn't get cleaned out.
>>>
>>> Once I removed all the "updated" libraries and did a fresh install,
>>> things seem to be OK.
>>> Thanks to P.D. and D. M. for your suggestions.
>>>
>>> On 5/31/25 4:42 AM, peter dalgaard wrote:
>>>> Works fine here.
>>>>
>>>> Did you do something with your packages after installing?
>>>>
>>>> Reason that I ask: A number of people have been shooting themselves in the foot by copying old package directories over new ones - including the non-CRAN system packages that update.packages() can't fix.
>>>>
>>>> Or, if you have packages in a separate non-versioned library, then you may need update.packages(checkBuilt=TRUE).
>>>>
>>>> -pd
>>>>
>>>>> On 30 May 2025, at 23.00, Carl Witthoft <carl using witthoft.com> wrote:
>>>>>
>>>>> I suspect this is something simple I forgot to update or delete, but:
>>>>> iMac x86 recently upgraded to 15.5 Sequoia.
>>>>>
>>>>> I installed R 4.5 from CRAN.
>>>>> then tried to install the "gmp" library and:
>>>>>
>>>>> Error: package or namespace load failed for ‘gmp’ in get(Info[i, 1], envir = env):
>>>>> lazy-load database '/Library/Frameworks/R.framework/Versions/4.5-x86_64/Resources/library/gmp/R/gmp.rdb' is corrupt
>>>>> In addition: Warning message:
>>>>> In get(Info[i, 1], envir = env) :
>>>>> internal error 1 in R_decompress1 with libdeflate
>>>>>
>>>>> Then,
>>>>> library(ggplot2)
>>>>> Error: package or namespace load failed for ‘ggplot2’ in dyn.load(file, DLLpath = DLLpath, ...):
>>>>> unable to load shared object '/Library/Frameworks/R.framework/Versions/4.5-x86_64/Resources/library/glue/libs/glue.so':
>>>>> dlopen(/Library/Frameworks/R.framework/Versions/4.5-x86_64/Resources/library/glue/libs/glue.so, 0x0006): Library not loaded: /Library/Frameworks/R.framework/Versions/4.4-x86_64/Resources/lib/libR.dylib
>>>>> Referenced from: <E900C54A-FB6D-371D-B50B-512EB8D2F868> /Library/Frameworks/R.framework/Versions/4.5-x86_64/Resources/library/glue/libs/glue.so
>>>>> Reason: tried: '/Library/Frameworks/R.framework/Versions/4.4-x86_64/Resources/lib/libR.dylib' (no such file), '/System/Volumes/Preboot/Cryptexes/OS/Library/Frameworks/R.framework/Versions/4.4-x86_64/Resources/lib/libR.dylib' (no such file), '/Library/Frameworks/R.framework/Versions/4.4-x86_64/Resources/lib/libR.dylib' (no such file), '/usr/local/lib/libR.dylib' (no such file), '/usr/lib/libR.dylib' (no such file, not in dy
>>>>>
>>>>> So I need to clear something out that's pointing to the 4.4 paths. But aged me forgets what that is.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Carl Witthoft
>>>>> personal: carl using witthoft.com
>>>>> The Witthoft Group, Consulting
>>>>> https://witthoftgroup.weebly.com/
>>>>>
>>>>> _______________________________________________
>>>>> R-SIG-Mac mailing list
>>>>> R-SIG-Mac using r-project.org
>>>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>>>>
>>>
>>
>
--
Sent from my phone. Please excuse my brevity.
More information about the R-SIG-Mac
mailing list