On 28 February 2025 00:23:59 GMT-06:00, Qian Yun <oldk1...@gmail.com> wrote:
>On 2/28/25 1:09 PM, Waldek Hebisch wrote:
>>
>> On Linux SBCL generated executables does not follow normal ELF
>> rules, in particular 'strip' breaks them and tools intended to
>> compare binaries may find false positives. Practical solution
>> is to block such tools (which I assume HOMEBREW_NO_INSTALL_CLEANUP=1
>
>Yes, on macOS the failing step is "codesign", and homebrew
>consider the installation has failed and do the uninstall
>automatically.
This sounds like a possible Homebrew bug.
Did you try to file a bug report with them?
They are usually quick to respond.
> The env var prevents the uninstall step.
>
>Only one thread about it on SBCL mail lists:
>
>https://sourceforge.net/p/sbcl/mailman/sbcl-devel/thread/513947F9-2DFA-4653-BF89-34AC0E4F855B%40gmail.com/#msg37368303
>
>> is doing). Proper solution would wrap core into an object
>> file and link with SBCL "loader" (IIUC CMU CL folks did this,
>> but it was not ported to SBCL).
>>
>> We had code to use wrapper script, but our code is simpler
>> without it.
>>
>
>I guess the wrapper script would be similar to the current
>ECL approach? If the patch is not complicated, will you
>accept it?
>
>- Qian
>
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to fricas-devel+unsubscr...@googlegroups.com.
To view this discussion visit
https://groups.google.com/d/msgid/fricas-devel/B66A748D-E3CD-4FD1-B5AC-3E383F82E520%40gmail.com.
