Fix lldb debug scripts #13048
Fix lldb debug scripts #13048
Conversation
|
It's not really working for me. After a clean and rebuild: $ echo '"string".hash' > test.rb
$ make -C target/debug lldb
compiling ../../version.c
linking miniruby
lldb -o 'command script import -r ../../misc/lldb_cruby.py' miniruby -- ../../test.rb
(lldb) target create "miniruby"
Current executable set to '/Users/etienne/src/github.com/ruby/ruby/target/debug/miniruby' (arm64).
(lldb) settings set -- target.run-args "../../test.rb"
(lldb) command script import -r ../../misc/lldb_cruby.py
lldb scripts for ruby has been installed.
(lldb) b rb_str_hash_m Breakpoint 1: where = miniruby`rb_str_hash_m + 16 at string.c:4086:35, address = 0x0000000100304c60
(lldb) r
Process 86318 launched: '/Users/etienne/src/github.com/ruby/ruby/target/debug/miniruby' (arm64)
Process 86318 stopped
* thread #1, queue = 'com.apple.main-thread', stop reason = breakpoint 1.1
frame #0: 0x0000000100304c60 miniruby`rb_str_hash_m(str=4751833240) at string.c:4086:35
4083 static VALUE
4084 rb_str_hash_m(VALUE str)
4085 {
-> 4086 st_index_t hval = rb_str_hash(str);
4087 return ST2FIX(hval);
4088 }
4089
Target 0: (miniruby) stopped.
(lldb) rp str
bits: [ ]
T_NONE: No value
(lldb) ^DIt's better than main which is crashing but not a whole lot. |
|
ok I had checked |
|
That's weird I just tested it and How is your Ruby configured? This is how I'm configuring And my lldb output: |
|
I used something like this: ../../configure --prefix=/Users/etienne/.ruby/debug --cache-file=../../config.cache --disable-shared --disable-install-doc --with-ext=openssl,psych,json,+ CFLAGS=-march=native --with-libyaml-dir=/opt/homebrew/opt/libyaml --with-openssl-dir=/opt/homebrew/opt/openssl@3 --with-gmp-dir=/opt/homebrew/opt/gmp --enable-debug-env 'optflags=-O0 -fno-omit-frame-pointer'I removed the CFLAGS and get the same results. The debugging configure flags come from: ruby/doc/contributing/building_ruby.md Line 257 in 36daa86 I tried with |
|
Can you go back to the commit before 0350290 and check that it works as expected? |
|
Also when I run lldb I'm using |
Arg sorry indeed I can't even get something to work in that case. Something must be wrong with my environment, I've used your configure options on a clean repo and I still get the same behavior. I tried on multiple computers (under macOS) too. I did just manage to make it work under lima using clang, though 👍 It does solve the issue in that case. cc @byroot |
In ruby#13008 `RVALUE` was removed without replacement. This means the lldb scripts that relied on `RVALUE` stopped working. I updated the ones that were using it just for the bytesize to use `slot_size` and then round to the nearest power of 40. We can't use `slot_size` directly because in debug mode it's `48` but `RVALUE` is `40` bytes. For the `as_type` method, I updated it to check the type. It's only used for `bignum` and `array` so that's a simple change. Lastly, for the `dump_page` method I replaced it with `struct free_slot` since that's looking at the freelist. `struct RVALUE` has been removed from all the scripts and I verified that `rp` is fixed. I'm not confident the `dump_page` method is fixed, the freelist looks off, but for now this gets us closer.
36daa86 to
b0d2192
Compare
✅ All Tests passed!✖️no tests failed ✔️61930 tests passed(3 flakes) |
In #13008
RVALUEwas removed without replacement. This means the lldb scripts that relied onRVALUEstopped working.I updated the ones that were using it just for the bytesize to use
slot_sizeand then round to the nearest power of 40. We can't useslot_sizedirectly because in debug mode it's48butRVALUEis40bytes.For the
as_typemethod, I updated it to check the type. It's only used forbignumandarrayso that's a simple change.Lastly, for the
dump_pagemethod I replaced it withstruct free_slotsince that's looking at the freelist.struct RVALUEhas been removed from all the scripts and I verified thatrpis fixed. I'm not confident thedump_pagemethod is fixed, the freelist looks off, but for now this gets us closer.