On 13/12/2018 19:32, Leonard Mosescu wrote:
What's the consensus?
Personally I think that, even considering the potential issue that Paval
pointed out, the "target symbols add ..." is the most conservative
approach in terms of introducing new behavior. I'm fine with the current
directory lookup as well (the original change) since it's consistent
with DWARF lookup.
Yes, but it also regresses existing functionality. Now if I do something
completely nonsensical like:
(lldb) target create "/bin/ls"
Current executable set to '/bin/ls' (x86_64).
(lldb) target symbols add -s /bin/ls /tmp/a.txt
error: symbol file '/tmp/a.txt' does not match any existing module
lldb will print a nice error for me. If I remove the safeguards like you
did in your patch, it turns into this:
(lldb) target create "/bin/ls"
Current executable set to '/bin/ls' (x86_64).
(lldb) target symbols add -s /bin/ls /tmp/a.txt
symbol file '/tmp/a.txt' has been added to '/bin/ls'
which is a blatant lie, because /bin/ls will continue to use symbols
from the object file.
_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits