wim@es.ele.tue.nl (Wim Philipsen) (03/14/91)
At about the same time we recieved out at&t C++ front-end with the cc6.7, and on an other tape 6.8. Did anybody have problems with the combination c++ and cc6.8? -- __ Wim J. M. Philipsen | Phone: +(31)40-473352 Eindhoven University of Technology, Dept. of Electr. Eng. (ES/EH 7.34) P.O. Box 513, 5600 MB Eindhoven, The Netherlands Email: wim@es.ele.tue.nl
perty@nmpcad.se (Per Lundholm) (03/15/91)
In article <WIM.91Mar14122747@eutes3.es.ele.tue.nl> wim@es.ele.tue.nl (Wim Philipsen) writes: > At about the same time we recieved out at&t C++ front-end with the cc6.7, and > on an other tape 6.8. Did anybody have problems with the combination c++ and > cc6.8? Same thing here. I have C++ code that runs if I use cc6.7 but crashes with a segmentation fault if I use cc6.8. It happens when calling a virtual function. I haven't been able to track it down enough to make up an APR. -- -------------------------------------------------------------------------------- Per.Lundholm@nmpcad.se ,____ NMP-CAD / / Swedish Institute of Microelectronics /_____/ ,____ ,__ / P.O. Box 1084 / /____/ / / -/- \ / S-164 21 Kista / (___ / (_ \/ __/ "My mind is a potato field..." /Zippy
vasta@apollo.HP.COM (John Vasta) (03/20/91)
In article <WIM.91Mar14122747@eutes3.es.ele.tue.nl> wim@es.ele.tue.nl (Wim Philipsen) writes: > >At about the same time we recieved out at&t C++ front-end with the cc6.7, and >on an other tape 6.8. Did anybody have problems with the combination c++ and >cc6.8? From the release notes for C 6.8 (section 1.10: C++ compatibility): To ensure compatibility between Domain/C Version 6.8 and either version of Domain/C++, compile your C++ source files with the -W0,-ncompress option. If you do not use -W0,-ncompress, constructors and destructors for static objects will not execute correctly. To ensure compatibility between Domain/C Version 6.8 and Domain/C++ Version 2.0.0, do not declare any virtual function inline. If you declare a virtual function inline, a bug in the C++ translator causes it to generate incorrect C code, and the C compiler issues an error message. Since the C++ translator ordinarily does not perform inline expansion of virtual functions in any case, this restriction usually has no effect on the quality of your code. Domain/C++ 2.1 (which should be shipping now) supports the use of compressed data, so if you get it you can stop turning compression off. The problem with inline virtual functions still exists, however. John Vasta Hewlett-Packard Apollo Systems Division vasta@apollo.hp.com M.S. CHR-03-DW (508) 256-6600 x5978 300 Apollo Drive, Chelmsford, MA 01824 UUCP: {decwrl!decvax, mit-eddie, attunix}!apollo!vasta