idb2pat fix bugs

[HongThatCong]

Hi William, I am TQN, the old author of idb2sig plugin.
I need to update my plugin, so I used your script to test the result and compared the resulting .pat files with IDA flair tool.
After some work, I found some bugs in the old plugin code.
So I have fix those bugs in your code. But it still have a small bug:
find_ref_loc will return BADADDR for the following sequence:

48 8D 05 29 06 00 00         lea     rax, `vtable for'loader_failure_t+10h

with:

• dref_from = 0xC40
• ref = 0x619

But in byte pattern is 0x0629 (29 06) because of +10h
Hope you have a time and fix it.
Thanks and best regards.
My code:
https://github.com/HongThatCong/idb2sig

[williballenthin]

hey @HongThatCong thanks for reporting this issue! I'll look into the bug and try to make a fix here. thanks for your hard work over the years that enabled me to learn how FLIRT works.

[HongThatCong]

Thanks Willi,
My friend have fixed this bug, and I recently uploaded.
But still have a bug. It will failed on ARM. On x86 and x64, it run OK. The result match with IDA flair tool.
I will add the form for user choose the type of functions which created pat ???

[HongThatCong]

Hi Willam.
I have fixed all known bugs.
Can you take a moment to test it.
I hope it will be updated to your code.
Sorry for my bad Python code and my English. I am Vietnamese.
Thanks and best regards...

[williballenthin]

I'm glad we can communicate! Thank you very much for reaching out here :-)

I will take a look at your fork and try to merge the changes in here. I'll update this issue when this is done. Thank you!