jjdredd
jjdredd
Looks like process receives SIGPIPE when trying to write to stdout. Usually this happens the second time when I click on "preferences".
I've managed to work around this by ignoring the SIGPIPE, but I still don't know why this bug appears.
Using option `--no-output-redirect` seems to fix the issue.
No, this is a separate issue.
Odd, I haven't encountered any bugs in the builder itself! Are you decompiling stripped bytecode? It could be a good idea to share the .luac file itself.
So, some closed-source version of luajit was used, right? Having source or docs it wouldn't be too hard to include new bc in ljd (even I could try). What is...
main.luc file header provided by @Bluetiger indicates that it's luajit 2.1 I successfully fed main.luc to luajit 2.1 (git 7f01300) and disassembled the file, so this bytecode is v2.1. Providing...
looks like we have a few new opcodes: ``` diff diff --git a/src/lj_bc.h b/src/lj_bc.h index 7436fab..64c1bcd 100644 --- a/src/lj_bc.h +++ b/src/lj_bc.h @@ -89,6 +89,8 @@ _(ISFC, dst, ___, var, ___)...
Adding new instructions to the ljd parser is not complicated. The hard thing is rechecking header format and adding those instructions for use in ast builder and unwarper. In addition...
How are you planning to add new opcodes without description? One way I see is to compile a bunch of code, seek for those opcodes in bytecode dump and try...