Ok, next: what do the unexpanded sccs keywords on the
#pragma ident
lines look like, i.e. since I'm using usr/src/lib/libc/port/gen/strlen.c
as a starting point, what should I make that line look like?
Since I tried to keep the resemblance between strlen.c and strnlen.c clear,
does the AT&T copyright notice stay in?
While I do have a copy of the most recent source downloaded (lots of changes
since the last one I downloaded, like how the mapfiles are set up!), I don't
have
the space {, experience, patience, ...} to attempt a full build, nor a machine
to
trash by attempting to install it. I hope that isn't necessary.
I did a little test program (not meant to fit into any sort of general test
infrastructure)
that gives strnlen() a workout, to test it with different ( <= max appropriate )
values of maxlen, check for correct results under termination both by NUL
character
and by maxlen, etc. By mmap'ing two adjacent anonymous, private pages, with
the 2nd having no privs, and the first read-only while strnlen() is called, it
also
checks that strnlen() does not look at memory beyond what it should and
(although it should be obvious) does not write to what is supposed to be
a read-only string. That is, the first arg to strnlen() is an arbitrary fairly
short area at the end of the first of those two pages. If that test program
were
a bit more generalized, it might be appropriate for other functions that treat
an area of memory as (your choice) read/write or read-only, and shouldn't run
off
the end. (It does indeed get a SEGV if one purposely messes up and runs into
the 2nd page!)
I saw Casper's post about mapfile-vers; would that, and the addition to
<string.h>,
be preferred in the form of context diffs, or what?
How about an addition to usr/src/lib/libc/port/llib-lc ? If so, what
visibility ?
I gather I'd have to fax in an SCA
(http://www.opensolaris.org/os/about/sun_contributor_agreement/); I can do that.
What other paperwork or other activity does there need to be on my part to get
this done?
That is, as long as I'm not in for much more than the above plus contributing
to the
justification for the ARC submission and suchlike minor stuff, I'm game for
this.
I suppose you saw the email from Dan Groves that said he put in CR 6478299,
whatever that means.
This message posted from opensolaris.org
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
