home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!att-out!rutgers!gatech!purdue!mentor.cc.purdue.edu!noose.ecn.purdue.edu!ecn.purdue.edu!kudva
- From: kudva@ecn.purdue.edu (Gautham K. Kudva)
- Newsgroups: comp.lang.c++
- Subject: SUMMARY: C++ 3.0 for Suns
- Message-ID: <1992Sep4.172502.18741@noose.ecn.purdue.edu>
- Date: 4 Sep 92 17:25:02 GMT
- Sender: news@noose.ecn.purdue.edu (USENET news)
- Distribution: usa
- Organization: Purdue University Engineering Computer Network
- Lines: 136
-
- A few days ago, I posted a request for info. on cfront 3.0 compliant
- C++ compilers for Suns. A few people expressed interest in the responses,
- so here's the list of replies I received:
-
- Thanks to the following people for their comments:
-
- Kurt Kanaskie
- Amrit Prem
- Thomas Wolf
- Ed Grossman
- Sue Lindsey
- Doug Shaker
- Shirley Kumamoto
- Joe Wilson
-
- We've pretty much decided on USL C++ 3.0. Their edu. pricing is pretty good,
- and we haven't heard any adverse reactions from other quarters either...
-
- Once again thanks to everyone who responded.
-
- Gautham
- kudva@ecn.purdue.edu
-
- -------------------------------------------------------------------------
-
- My original post was:
-
- > Hi,
- > Our research group has recently started using C++.
- > We do most of our development work on Sun SPARC workstations.
- > So far we've been going with cfront 2.1, and we'd like to upgrade
- > to C++ version 3. Our primary reason for upgrading is Template
- > support. Does anyone have experience with any cfront compliant
- > C++ compilers with template support for Suns? In particular how
- > about USL C++ 3.0? Any comments or suggestions are greatly
- > appreciated.
- >
-
- ========================================================================
-
- Gautham,
-
- We have been using USL C++ 2.1 for over a year now. We recently switched
- to USL C++ 3.0. The only noticable difference is the extra checking for
- "jump beyond initializer" in case statements where a local variable is
- defined.
-
- ========================================================================
-
- We use Object Store C++ here in house. I've used the template facility for
- a few simple things. The main problem with their templates is that
-
- a. Function templates are not yet supported.
-
- b. If the template requires non-inline code to be written, then
- the module where the code is defined in must instantiate all
- of the template classes that will be used in the application.
-
- The compiler is basically a hack of cfront, so you get cfront compatibility
- with templates thrown in. (This means you also get cfront's bugs...)
-
- ========================================================================
-
- Our group has been using cfront 3.0 for a couple of months now and have had
- no problems...just a little slower compiles on C++ modules including templates.
- We are running it on Sun SPARC workstations.
-
- ========================================================================
-
- I've been using 3.0/3.01 for about 8 months on a SPARC.
- The template mechanism has a fair number of bugs if you
- try to use it for anything that's not entirely straightforward
- and the template specialization mechanism (ie providing
- a special version of a tempalted class) doesn't really
- work at all. So far, however, I haven't found anything
- that's totally stymied my work. I've not yet seen anything
- better, although I'm thinking of checking out GNU's version
- after it settles down a bit. I've been using templates to
- provide sets and smart pointers (and sets of smart pointers)
- and I seem to be constantly pushing the envelope of what the
- compiler will accept.
-
- ========================================================================
-
- TauMetric Corporation now develops, maintains, and sells the Oregon C++ compiler
- and debugger. These compilers run under VAX/VMS, DECstation/Ultrix,
- SPARC, and 6800x0 workstations like Sun-3, or HP 9000/300, and 386 machines
- running Interactive, UHC, or AT&T Unix/386 3.2 or 4.0.
-
- The Oregon C++ compiler is a true compiler, it compiles C++ code to object code.
- It features three switches to allow the user to compile C++, ANSI C, or
- K&R style C programs. A source level debugger is included with each compiler.
- The debugger is specifically written for C++, it understands multiple
- inheritence and classes.
-
- We will be releasing our 3.0 Oregon C++ compiler for SPARC later this year.
- It is our own compiler, not cfront.
-
- If you have further questions you can contact me at 619-697-7674, or
- send me questions via e-mail at: sue@taumet.com.
-
- ========================================================================
-
- The only released cfront 3.0 for Sparcs that I know about is
- Solbourne's cfront. We sell it if you are interested. Send
- email to info@qualix.com for more information.
-
- - Doug Shaker
-
- ========================================================================
-
- The Qualix Group provides another commercial 3.0 cfront (from
- Solbourne Computing) for SPARC stations. Its a full function
- version of the AT&T C++ Language System Release 3 and translates
- into C source code. If you would like more information,
- let me know.
-
- Regards,
-
- Shirley Kumamoto
- Product Manager
- Qualix Group
- 1900 S. Norfolk St. Suite 224
- San Mateo, CA 94403-1151
- voice: (415) 572-0200
- fax: (415) 572-1300
- email: shirleyk@qualix.com
-
- ========================================================================
-
- Gautham:
- cfront 3.0.1 supports template classes. We're using them heavily.
- Sun will be releasing this soon. It's available from ATT USL today.
- (800-828-unix).
-
- ========================================================================
-