kaiser@ananke.stgt.sub.org (Andreas Kaiser) (09/09/90)
#if 0
The following example should print an error message, because
the non-static member <i> should not be accessible within the
static member function <f>. Instead I got bad code (in this
example) or an invalid fixup record (in a larger program).
Compiled with ZTC++ 2.12:
ztc -c err_24.cpp
#endif /*---------------------------------------------------------*/
struct S {
int i;
static void f();
};
void S::f() { i = 0; }
main() { }
#if 0 /*----------------------------------------------------------*/
[....]
public _f__1SNv,_main
extrn _S <--------
[....]
_f__1SNv:
mov _S,0 <--------
ret
[....]
#endif
Gruss, Andreas
---
--
:::::::::::::::::::: Smart: kaiser@ananke.stgt.sub.org
:: Andreas Kaiser :: Bang: ...!smurf.sub.org!nadia!ananke!kaiser
:::::::::::::::::::: Fido: 2:507/18.7206 & 2:509/5.2512joe@proto.COM (Joe Huffman) (09/10/90)
In article <17.26EA603B@ananke.stgt.sub.org>, kaiser@ananke.stgt.sub.org (Andreas Kaiser) writes: [...Bug report for Zortech C++ compiler...] You may mail your bug reports to the engineers, tech support, and tech writers by sending them to 'zortech-bugs@proto.com' or 'uunet!proto!zortech-bugs'. Posting them on the net is probably inappropriate most of the time. Thanks for taking the time to report it though. I have sent it on to the appropriate people. -- joe@proto.com FAX: 208-263-8772
daniel@terra.ucsc.edu (Daniel Edelson) (09/11/90)
In article <1430@proto.COM> joe@proto.COM (Joe Huffman) writes: >In article <17.26EA603B@ananke.stgt.sub.org>, kaiser@ananke.stgt.sub.org >(Andreas Kaiser) writes: > >[...Bug report for Zortech C++ compiler...] > >You may mail your bug reports to... >.... Posting them on the net is probably >inappropriate most of the time. I disagree. We in the community benefit from this information, until of course it becomes tedious or redundant. You probably just would rather not have your bugs reported widely in the expectation that this will present a negative image. But if bugs in all the products are reported, then no vendor is treated unfairly. And this benefits the group of compiler users, and encourages compiler vendors to eliminate important bugs before they set out to implement esoteric features. Several times people have posted bugs that pointed out either problems in their coding or in the compiler, such as the malloc discussion of the past week or two. This is worthwhile. A huge number of bugs in Cfront 2.0, as well as several g++ bugs have been posted by Ron Guilamette. We have also heard about some bugs in Zortech and Borland's products, though not as many because the external testing of those products has not been (I believe) performed by rfg. The companies with C++ validation tests are understandably reluctant to post their results. Such comments may highlight aspects of the language that are particularly difficult to implement, as if Bjarne would propose anything hard to implement ;+). Daniel Edelson daniel@cis.ucsc.edu, or ...!sun!practic!peren!daniel
bright@Data-IO.COM (Walter Bright) (09/11/90)
In article <6744@darkstar.ucsc.edu> daniel@cis.ucsc.edu (Daniel Edelson) writes: <In article <1430@proto.COM> joe@proto.COM (Joe Huffman) writes: <<In article <17.26EA603B@ananke.stgt.sub.org>, kaiser@ananke.stgt.sub.org <<[...Bug report for Zortech C++ compiler...] <<You may mail your bug reports to... <<.... Posting them on the net is probably <<inappropriate most of the time. <I disagree. We in the community benefit from this information, <until of course it becomes tedious or redundant. Actually, the trouble is that we (Zortech) don't always get the bug reports if they are posted to various networks. For instance, we don't know about the other 23. I know Andreas posted some of them to Fido, but my Fido connection is flaky and I didn't get even a quarter of them. People should at least email a copy of it directly to us, so we can be sure and get the problems resolved. There is a mailing list for Zortech bug reports, and people interested in them can subscribe, rather than boring everyone on the net about it.