[ ant-contrib-Bugs-1718197 ] Should Sun CC bid on .c++ files

Bugs item #1718197, was opened at 2007-05-13 16:22
Message generated for change (Comment added) made by darius42
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=416920&aid=1718197&group_id=36177

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: cpptasks
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: David Haney (darius42)
Assigned to: Curt Arnold (carnold)
Summary: Should Sun CC bid on .c++ files

Initial Comment:
This issue was originally raised in Bug 1715475:

The behavior of the Forte CC compiler changed with Forte 6.1 (Sunpro 5.2).  Prior to Forte 6.1, the .c++ extension was not supported (and the compiler would reject it as an unknown file type).  With 6.1 and later, the .c++ extension is supported.  So should the Forte adapter bid on .c++ files (since they are currently supported) or not (since some compiler versions won't be able to handle them).

Initial response from carnold:

Maybe a half-hearted bid (something like DEFAULT_PROCESS_BID/2), so if no other compiler makes a full bid, then it will get a chance on the file.  Better for a compiler that might be able to process it gets it than the linker (which gets all unbid files).

----------------------------------------------------------------------

>Comment By: David Haney (darius42)
Date: 2007-05-13 17:13

Message:
Logged In: YES 
user_id=1140024
Originator: YES

Would it make sense to query the compiler version and use that to
determine whether .c++ files should be bid on?  For CC, we should be able
to use "CC -V" in order to query the version.  Once we have that, we could
bid on .c++ if there version is greater than or equal to c5.2.  That way
the bid would be appropriate to the platform.  Something similar could be
used for bug 1690945 in order to decide which flags to use when invoking
the cl.exe compiler.

It looks like we already query the version for the gcc compiler, so
there's some precedence for this approach.  In other places however it
looks we've avoided querying the compiler due to hangs, however I'm not
sure off hand what would cause that.

Querying the compiler does complicate testing, since any verification of
this behavior would have to be executed on machines that have the compiler
installed, and would have to tested against machines with different
versions of the compiler installed.

Another alternative would be to have separate enum for sunpro52.  This was
mentioned in 1690945 for addressing the msvc8 differences, and may be
appropriate here.  

Version detection seems like the cleanest solution (if it works) for
users, since they won't have to worry about what version of the compiler
they have installed when setting up their ant configurations.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=416920&aid=1718197&group_id=36177

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/