GCC Moves Forward With Conversion To C++

[Mystro256]

Hey then we all should be switching to Mono right?

Haha, I can't even read that with a straight face.

[frantaylor]

In the distant past, this would have met with rolling eyes, because all the other vendors' C++ compilers were so crappy that you wouldn't be able to bootstrap gcc. Haha the good old days of bootstrapping gcc onto Solaris 2.6 so we could run gimp on the ultrasparc with the big monitor.

I've had some good hearty laughs at the error messages, as I've attempted to port "standard" C++ code to Unixes like HPUX and Tru64.

In the end I had to rewrite it in C, because that was the only subset of C++ that was actually portable.

If your C code is already getting compiled as C++ anyway, there's certainly no harm in grabbing a few of the more useful bits. The whole C/C++ thing is a continuum, so you can use as much or as little of C++ as you like.

Bad C++ code comes from bad C++ programmers, not from C++ compilers.

Yeah I've implemented "virtual functions" in C, but it's truly a pain in the butt, I'll take C++ virtual functions anytime.

[elanthis]

Kudos to him for copying UNIX verbatim but is there anybody on the planet more opinionated than him?

Hello.

[bug!]

Why C++ and not Python or even Lua? Those higher level languages allow for more abstraction. Python, for one, has an object oriented model with less C-like hackery intermixed. Embedding one of them into the core would probably benefit more.

Let's see if GCC devs dig their own grave or manage to stay afloat. I'm kind of skeptical about the latter without reimplementing toolchain in C++ from scratch (look at LLVM/Clang).

[bug1]

Why C++ and not Python or even Lua? Those higher level languages allow for more abstraction. Python, for one, has an object oriented model with less C-like hackery intermixed. Embedding one of them into the core would probably benefit more.

Let's see if GCC devs dig their own grave or manage to stay afloat. I'm kind of skeptical about the latter without reimplementing toolchain in C++ from scratch (look at LLVM/Clang).

Great alias's think alike !

[TheBlackCat]

The reasons they cite being

• C++ is a standardized, well known, popular language.
• C++ is nearly a superset of C90 used in GCC.
• The C subset of C++ is just as efficient as C.
• C++ supports cleaner code in several significant cases.
• C++ makes it easier to write and enforce cleaner interfaces.
• C++ never requires uglier code.
• C++ is not a panacea but it is an improvement.

None of which are technical benefits, GCC are choosing to compromise on quality for the percieved convenience of its implementation.

The role of a programmer is to translate real world problems into the realm of the machine, this change clearly doesnt reflect that understanding. You would think, as compiler developers they would understand its purpose better.

You should read the article that is quoted from. They list a number of tasks they intend to accomplish with the conversion, and each task provides a description of how the switch to C++ will have specific, technical benefits to GCC. If you want to argue against the benefits of the chance, it would be better to argue against the specifics in the original article, rather than the summary quoted by somebody not involved in the project.

[TheBlackCat]

Why C++ and not Python or even Lua? Those higher level languages allow for more abstraction. Python, for one, has an object oriented model with less C-like hackery intermixed. Embedding one of them into the core would probably benefit more.

Obvious troll is obvious, but I think it would have been better if they followed the PEP system python uses. The proposal article provides a good overview of the logic behind the change, but it lacks the detail and rigor of a PEP.

Let's see if GCC devs dig their own grave or manage to stay afloat. I'm kind of skeptical about the latter without reimplementing toolchain in C++ from scratch (look at LLVM/Clang).

You do realize that C++ is largely a superset of C, right? The first step, as they said, was making GCC build with C++ instead of C (which is already done). After that, they can use as many or as few C++ features as they want, picking only those features that will provide a technical benefit (the article, as I said, explains in some details what the specific features they wish to implement are and what their benefits are).

[curaga]

They can write in whatever language they prefer. I only lament the fact that future GCCs will take about 20x longer to build due to this.

[mirv]

They can write in whatever language they prefer. I only lament the fact that future GCCs will take about 20x longer to build due to this.

Why would it take longer?
Regarding putting C++ into gcc, there's been talk of it for a long time, and their general guidelines are to only use something if it makes things "better". There are some things for which C++ is better suited than C, and the idea is to use C++ for those cases. When I say use C++, of course, I mean "something in C++ that's not in C".

[Awesomeness]

This is a bad thing, right? C++ is such a horrible and ugly language.

As opposed to beautiful plain C?