BIND 10 #3213: BIND 10 doesn't compile/work on Mac OS X 10.9

[BIND 10 Development]

#3213: BIND 10 doesn't compile/work on Mac OS X 10.9
-------------------------------------+-------------------------------------
            Reporter:  jinmei        |                        Owner:
                Type:  defect        |  jinmei
            Priority:  medium        |                       Status:
           Component:  build system  |  reviewing
            Keywords:                |                    Milestone:
           Sensitive:  0             |  Sprint-20131015
         Sub-Project:  Core          |                   Resolution:
Estimated Difficulty:  0             |                 CVSS Scoring:
         Total Hours:  0             |              Defect Severity:  N/A
                                     |  Feature Depending on Ticket:
                                     |          Add Hours to Ticket:  0
                                     |                    Internal?:  0
-------------------------------------+-------------------------------------

Comment (by jinmei):

 Mukund, thanks for the review.

 Replying to [comment:4 muks]:

 > Here is my review of this branch. Will you make these changes (if you
 > agree) to the branch or do we have to make them?

 I've created a new branch named 'trac3213' for the suggested rebase,
 and addressed the comment in that branch.

 > * There is a compile issue on Fedora 19 with the GCC 4.8.2 compiler:
 > {{{
 > data.cc: In function 'int isc::data::{anonymous}::skipTo(std::istream&,
 const string&, int&, int&, const char*, const char*)':
 > data.cc:307:1: error: control reaches end of non-void function [-Werror
 =return-type]
 > }}}

 Actually, that happened on my Mac, too.  I don't know why I didn't see
 it before, but anyway, addressed it.

 > * There were clang++ fixes that were merged to `master` branch recently
 >   (to make it compile with clang++ 3.3 rc3 on Fedora 19). Please can you
 >   rebase this branch on `master`? The
 >   `-Wno-tautological-constant-out-of-range-compare` may be unnecessary.

 Yes, it's now unnecessary.  I've reverted the corresponding commit of
 mine.

 > * In the `pthread_cond_destroy()` patch, it is testing using
 >   `/usr/bin/sw_vers` for a specific version. A long time ago, when
 >   working on a different ticket, I think you had advised me (maybe when
 >   working on a Solaris issue) that it is better to test for
 >   features/bugs directly and set configure variables based on those,
 >   instead of any system specific tools or versions. Would it make sense

 Yes, that's my general belief, and, I don't remember it was with you I
 remember I had that kind of discussion, so...

 >   to test whether `EBUSY` is returned in this case or not?

 ...I wish I could do this, but in this case it would be too much for
 something like configure.ac (or its helper m4 macro); we'd need to
 embed check code that uses multiple threads and run it.  So I think
 this is one of the exceptional cases where the OS/version based switch
 is acceptable as a compromise.  I've added some comments in
 configure.ac to explain that.

 I also made one additional change in masterload.{c,h}: I noticed
 parameter 'source' for the istream version is now effectively unused,
 and then we can easily consolidate the two versions.  So I did that,
 and updated the documentation about the now-unused parameter.

-- 
Ticket URL: <http://bind10.isc.org/ticket/3213#comment:6>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development