<div dir="ltr"><div dir="ltr"><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Sat, Jul 16, 2022 at 12:14 AM Rob Landley <<a href="mailto:rob@landley.net">rob@landley.net</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On 7/15/22 21:24, enh via Toybox wrote:<br>
> i saw you complaining about memory errors on your blog, talking about writing<br>
> your own infrastructure... why not just `export ASAN=1`?<br>
<br>
Ray Gardner emailed me identifying the line on wednesday (quote_filename()<br>
returns garbage so free(quoted_file) breaks).<br></blockquote><div><br></div><div>yeah, that's what asan said. my point was, for the low low cost of just putting `export ASAN=1` in your .bashrc, you wouldn't ever have to even ask. the increase in compile time is pretty small, and for things the size of toybox tests, the runtime cost is fine too. (my makefiles for my personal hobby projects don't even have a non-asan option.)</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
My complaint was more "Ze <strike>goggles</a> test suite tests, ze do nothink!"<br>
because the bug WASN'T in my current tree, it was in the previously checked in<br>
stuff that had passed the test suite before being checked in.<br></blockquote><div><br></div><div>aye, but that's an orthogonal problem. and the nice thing about running asan builds all the time is that you catch a bunch of stuff by accident. that's _why_ it's so great. (you don't even need a corruption/UAF bad enough that the "production" code would even crash... you can find subtle stuff with easy cases.)</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
Rob<br>
</blockquote></div></div>