From: Earnie B. <ear...@ya...> - 2002-11-28 12:41:29
|
Your presentation of analysis has persuaded me to change my vote to sjlj. Earnie. Wu Yongwei wrote: > I want to add that: > > * sjlj: You will have slower performance, but only in a try..catch block > (even if no exceptions are really thrown). You will have a smaller > executable size. > > * Dwarf2: You will have faster performance in a try..catch block in the > most probable case that exceptions are not thrown. But you will have a > bigger executable size even if you do NOT use exceptions at all. -- This > is one thing that I do not like about Dwarf2. > > So it is in some ways also a size-speed dilemma, as I understand it > (only in smaller applications, of course; in big applications you will > not note the size difference). Neither way is perfect, in that none of > them completely supports the C++ design principle that one needs not to > pay for a feature if one does not use it. IMHO, sjlj is at least better > in the case when one does not use C++ exceptions. But others may have > different opinions. > > It is cool if GCC can support both methods simultaneously by a > compile-time switch, but I doubt anyone will do it. > > Correct me if there are any false statements. I am not an expert in this > aspect. > > Best regards, > > Wu Yongwei > > --- Original Message from Earnie Boyd --- > > For the record, Danny didn't say that Dwarf2 was "unstable", Oscar said > that. The problem that Drawf2 has thst sjlj fixes is being able to > throw an exception in a CALLBACK function and have it caught be the > calling function. So the question isn't about stability it's about: > > [ ] Enhanced performance with not being able to throw exceptions in a > CALLBACK function. Dwarf2. > > [ ] Slower performance and being able to throw exceptions in a CALLBACK > function. sjlj. > > Earnie. > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Get the new Palm Tungsten T handheld. > Power & Color in a compact size! > http://ads.sourceforge.net/cgi-bin/redirect.pl?palm0002en > _______________________________________________ > MinGW-users mailing list > Min...@li... > > You may change your MinGW Account Options or unsubscribe at: > https://lists.sourceforge.net/lists/listinfo/mingw-users > |