Sigh. I am juggling too many things the past couple days. Rob
On 02/29/2016 11:50 PM, enh wrote: > i think you want this patch rather than the one you committed > (c1161cf09f5d9646b0e0c6ec504ed26a0623e691) --- doesn't yours leak fd? > > On Mon, Feb 29, 2016 at 4:25 PM, enh <e...@google.com> wrote: >> what's checked in now still doesn't work in the regular file case though... >> >> >> Subject: [PATCH] Fix file "cannot open". >> >> We don't want to _always_ fall through. >> --- >> toys/pending/file.c | 3 ++- >> 1 file changed, 2 insertions(+), 1 deletion(-) >> >> On Mon, Feb 29, 2016 at 1:15 PM, Rob Landley <r...@landley.net> wrote: >>> Already checked these two in before I saw this. (I checked in the >>> architecture name list redo last night because the changes to file were >>> turning into a rathole, but "I tested this and it worked" only means >>> something if you haven't changed it since. Sigh...) >>> >>> It should be working now. And I can make test cases that check /dev is a >>> directory, /dev/zero is a char device, /dev/loop0 is a block device (as >>> standard of one as we're likely to get anyway), mkfifo a fifo, ln -s >>> makes a symlink... >>> >>> Huh, >>> >>> $ file /usr/lib/firefox/firefox.sh >>> /usr/lib/firefox/firefox.sh: POSIX shell script, ASCII text executable >>> >>> According to posix that should reply "commands text". Once again, posix >>> is completely ignored in practice. >>> >>> Sigh, >>> >>> Rob >> >> >> >> -- >> Elliott Hughes - http://who/enh - http://jessies.org/~enh/ >> Android native code/tools questions? Mail me/drop by/add me as a reviewer. > > > _______________________________________________ Toybox mailing list Toybox@lists.landley.net http://lists.landley.net/listinfo.cgi/toybox-landley.net