Feedback-ID: i5b2c416c:Fastmail
MIME-Version: 1.0
Date: Tue, 18 Mar 2025 19:00:00 -0400
From: "Jim Pryor" <dubiousjim@gmail.com>
To: Sertonix <sertonix@posteo.net>, ~alpine/devel@lists.alpinelinux.org
Message-Id: <9b95d46b-88c3-4443-8db8-96b457209d7b@app.fastmail.com>
In-Reply-To: <8c748e1c-78a2-4073-9d19-32501607d7e9@app.fastmail.com>
References: <30674376-d675-4c1d-96ef-7afb88a0dabb@app.fastmail.com>
 <D8JPX4S6FPXS.1JHNGLY840E49@posteo.net>
 <8c748e1c-78a2-4073-9d19-32501607d7e9@app.fastmail.com>
Subject: Re: stack smashing protection?
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

I realized that it's not surprising that the alternative fixes I mentioned aren't working, since supplying them to the gcc build command isn't going to pass them on to the non-standard linker (the mpy_ld.py script) used in this case. And that tool doesn't have command-line options that look promising. I can hack its source directly though.

The missing symbol is found in the /usr/lib/libc.so file (well, that's a link, but it's found in its target). But that doesn't seem to be linked into the binary that's produced.

There is a /usr/lib/libc.a file. I can look into whether some *.o file extractable from that defines the symbol I'm missing...

-- 
dubiousjim@gmail.com