Zapcc Claims To Be A "Much Faster C++ Compiler

[Uqbar]

Published on the main Phoronix site at http://www.phoronix.com/scan.php?pag...-Compiler-Fast .

Then, my personal opinion is that the time you spend in compilations is less important (by faaar) that the quality of the machine code produced by the same.
Let's say you save 10% of the compilation time with this tool. If the result machine code runs even 1% slower than "that other slooow compiler"(tm), you did a bad job.
For every compilation I foresee million of program run.
I would focus on execution time and machine code effectiveness, just like GCC and LLVM are doing.

[haagch]

On their website

No link?

gcc drop-in replacement
clang is usually a plug-in replacement for gcc, clang maintains this feature.

Should have told someone to read their website before publishing.

I've looked around their website but didn't see anything: It will be closed source, right?

[caligula]

Seems like a silly way to waste your time. C and C++ are the only survivors in this age with totally crappy "module" systems. First, the syntax is total pain in the ass. Second, the system makes it insanely hard to do incremental compilation and modular programming. Even if the later standards would include some standard "import" system and modules, they wouldn't be first class modules and you'd still need to support old #include scheme for backwards compatibility.

[DarkCloud]

Seems like a silly way to waste your time. C and C++ are the only survivors in this age with totally crappy "module" systems. First, the syntax is total pain in the ass. Second, the system makes it insanely hard to do incremental compilation and modular programming. Even if the later standards would include some standard "import" system and modules, they wouldn't be first class modules and you'd still need to support old #include scheme for backwards compatibility.

I don't understand your argument here. If C and C++ are a waste what are you going to write your modern runtime languages in - Assembly ? And as far as modules goes, there are advantages for statically typed languages such as C and C++. Both C and C++ still very relevant for many kinds of applications.

[adler187]

No link?

Came here to post the same.

BTW Michael, the article isn't wired up to this comment thread correctly, so the "Add a comment" just takes you to the sub-forum for the Phoronix Articles. Oh, I see that you didn't create the forum post, Uqbar did - maybe that has something to do with it.

[gens]

don't people realize that steroids are a Bad Thing ?

Seems like a silly way to waste your time. C and C++ are the only survivors in this age with totally crappy "module" systems. First, the syntax is total pain in the ass. Second, the system makes it insanely hard to do incremental compilation and modular programming. Even if the later standards would include some standard "import" system and modules, they wouldn't be first class modules and you'd still need to support old #include scheme for backwards compatibility.

C != C++

C does not have "modules", it has "functions" and "data"

[peppercats]

Published on the main Phoronix site at http://www.phoronix.com/scan.php?pag...-Compiler-Fast .

Then, my personal opinion is that the time you spend in compilations is less important (by faaar) that the quality of the machine code produced by the same.
Let's say you save 10% of the compilation time with this tool. If the result machine code runs even 1% slower than "that other slooow compiler"(tm), you did a bad job.
For every compilation I foresee million of program run.
I would focus on execution time and machine code effectiveness, just like GCC and LLVM are doing.

Gee, I'm sure glad Clang isn't under the GPL so we don't get to enjoy this benefit that was built entirely upon work that wasn't their own.