I got distracted over the weekend, but lemme see if I can come up with a more elegant solution this evening before falling back to that plan.
Thanks, Rob On 05/07/2018 01:50 PM, enh wrote: > testing with `cat boot.img | strace file -` it looks like FSF > allocates a ~10MiB buffer and reads it all in. (rather than copying to > a temporary file.) > > happy to supply that patch instead if you prefer? > > On Sat, May 5, 2018 at 4:45 PM, Rob Landley <[email protected]> wrote: >> On 05/05/2018 06:19 PM, enh wrote: >>> AFK, but I'm using glibc too, from Debian testing. >> >> I found it, the stdin case is using the stat.length value to see how much to >> read but never did the stat, so it's uninitialized stack crap. When it's >> zero... >> >> Oops. Oddly enough, the compiler did not complain! >> >> Needs a design tweak to figure out what the length _should_ be in this case. >> Maybe that read-into-memory suggestion with like a 256k cap or something? >> Working on it... >> >> Rob > _______________________________________________ Toybox mailing list [email protected] http://lists.landley.net/listinfo.cgi/toybox-landley.net
