Misc. bug: llama-quantize clobbers input file + crashes when output file matches

[m18coppola]

Name and Version

version: 5029 (833e2b7)
built with cc (GCC) 13.2.0 for x86_64-unknown-linux-gnu

Operating systems

Linux

Which llama.cpp modules do you know to be affected?

llama-quantize

Command line

llama-quantize ./Llama-3.2-1B-Instruct-f16.gguf Llama-3.2-1B-Instruct-f16.gguf 7

Problem description & steps to reproduce

If the input and output files point to the same location, the program crashes:

[1]    12928 bus error  llama-quantize ./Llama-3.2-1B-Instruct-f16.gguf Llama-3.2-1B-Instruct-f16.ggu

This also clobbers the input file, rendering it nearly empty, and unusable:

$ du -sh ./Llama-3.2-1B-Instruct-f16.gguf
7.5M    ./Llama-3.2-1B-Instruct-f16.gguf

First Bad Commit

No response

Relevant log output