From -3650999904244509869 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: f78e5,e445fe05b100a0aa X-Google-Attributes: gidf78e5,public X-Google-ArrivalTime: 2002-08-23 03:30:02 PST Return-Path: X-Authentication-Warning: mulga.cs.mu.OZ.AU: fjh set sender to devnull@stump.algebra.com using -f Path: archiver1.google.com!news1.google.com!newsfeed.stanford.edu!comp-std-cpp-robomod!not-for-mail X-Robomod: STUMP, ichudov@algebra.com (Igor Chudov) Delivered-To: std-c++@ncar.ucar.edu From: "Ken Hagan" Newsgroups: comp.std.c++ Subject: Re: C++ exception handling Message-ID: <1030093822.25233.0.nnrp-10.3e31ffea@news.demon.co.uk> References: <20020802141846.59203.qmail@web9902.mail.yahoo.com> <_2S29.267$a01.33863850@newssvr13.news.prodigy.com> <1028797052.3258.0.nnrp-13.3e31ffea@news.demon.co.uk> <3D54EE8E.DCE1A08D@iobox.com> <1029750082.26799.0.nnrp-10.3e31ffea@news.demon.co.uk> <3D628B51.35E8C851@iobox.com> X-NNTP-Posting-Host: mailgate.thermoteknix.com:62.49.255.234 X-Trace: news.demon.co.uk 1030093822 nnrp-10:25233 NO-IDENT mailgate.thermoteknix.com:62.49.255.234 X-Complaints-To: abuse@demon.net X-Priority: 3 X-MSMail-Priority: Normal X-Newsreader: Microsoft Outlook Express 6.00.2600.0000 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Approved: Fergus Henderson , moderator of comp.std.c++ Date: Fri, 23 Aug 2002 05:25:42 CST Lines: 53 Xref: archiver1.google.com comp.std.c++:13356 "Sergey P. Derevyago" wrote... > Yes. It seems like we have different goals: > 1. I want to write exception safe code, so (statically checked) nothrow is the > solution, IMHO. > 2. While you want to futher extend those bra^H^H^H dead-born exception > specifications :) Actually my interest also lies with 1 rather than 2. It's just that I can't see how to achieve one without the other. I'm not prepared to declare all my template functions as throw(...), throw(), throw(std::exception) of anything else, because any particular ES will stop me using the template with certain types, at least usefully. Let me break down my chain of thought here. 1 I'm only interested in checking "nothrow" at a few places in my code. 2 The appropriate ES of other functions is just "whatever they can throw". 3 That depends on what they call. 4 For non-templates, "what they call" was fixed at compile-time, so the original ES proposal was workable. 5 For template functions it isn't fixed until instantiations are made; To check at compile-time, the ES would need to be a formula for "how to compute the actual ES for given template parameters". That's why I gave the impression of wanting to extend ES. In fact, I don't. I'd much prefer either of the following. [A] Force whatever tool does template instantiation, to generate an "actual ES" independently of what's in the source code. That can then be checked by for the entire program in one go. [B] Decide that the above is too much of a requirement for compiler vendors, but suitable for external tools like lint. In either case, the programmer would only need to write an ES where they wanted an actual check, and check failures would not prevent the build from being successful. With this kind of checking available, I can't see any need for the current run-time semantics, so I'd be inclined to ban^H^H^H deprecate them. --- [ comp.std.c++ is moderated. To submit articles, try just posting with ] [ your news-reader. If that fails, use mailto:std-c++@ncar.ucar.edu ] [ --- Please see the FAQ before posting. --- ] [ FAQ: http://www.jamesd.demon.co.uk/csc/faq.html ]