On Sep 1, 2007, at 3:23 PM, Mathias Gaunard wrote:
Chris Lattner wrote:
David's characterization is somewhat correct, but is also a bit simplistic. The LLVM project does certainly use templates (including partial specialization etc), but we prefer to keep this as an implementation detail in a library. Exposing "complex" template code through the public interface of a library generally make the library "scary" to those developers who don't consider themselves to be C++ gurus. This design point also reduces build time for the LLVM code itself.
From a rough sketch, I personally feel more scared by the pointers exposed everywhere -- which are one of the greatest sources of unsafety in C++ --, the non usage of RAII, the explicit casts and other unsafe things, the usage of downcasting everywhere etc.
This is an interesting assertion. In LLVM, we have had far more bugs and other problems due to various subtle iterator invalidation issues than due to any sort of pointer or memory lifetime problems. While they may be your personal fear, they seem to work well for a lot of code in practice.
I had read that LLVM was supposed to be "finely crafted C++" but it looks more like the usual object-oriented C++ from the nineties that you see everywhere and that contributed to make C++'s reputation of a dangerous and unmaintainnable language.
I'm sure you're aware that "the usual object-oriented C++ from the nineties" often is the best solution when you do have *dynamic* polymorphism - as you do when processing ASTs and other compiler IRs. Templates are great for solving problems involving static polymorphism, and we certainly do apply them to such problems. In any case, this hyperbole is wildly off-topic and not very useful. If you have specific suggestions on improvements that can be made to various LLVM APIs, please do bring it up on the appropriate LLVM lists - we're always very interested in improving the system.
I honestly don't understand how can people be "scared" to use templates, especially compiler writers, which should be able to understand how they work and what they do.
You need to talk to C programmers more often then. Not all compiler engineers are C++ front-end hackers after all. I'm not interested in artificially restricting the number of people who can contribute to LLVM. Using templates for the sake of using templates isn't interesting to me: using them where they add value is, and we continue to do so.
They don't add bloat unless you make them do so (especially with LLVM which should be able to remove all unused code and duplicates without any problem). On the contrary, they allow the design of thin layers for robust type safety as well as nice reusable wrappers.
Templates are a tool that can be used for good as well as for evil. When misused, templates, like any other powerful tool, can be very damaging to readability, compile time, and code size. When used appropriately, then can be quite beneficial. In any case, this discussion isn't very appropriate to the boost list. If you want to suggest ways to improve LLVM, please do so in the appropriate place: the LLVM list. Thanks! -Chris