commit 3a0ace595031beb4bdc8faf3606004347b41a7d0
parent 2e8eb91ca021069235d090bcb2aea543070772f8
Author: arg@www.suckless.org <unknown>
Date: Mon, 9 Mar 2009 15:36:34 +0000
update
Diffstat:
1 file changed, 29 insertions(+), 13 deletions(-)
diff --git a/www.suckless.org/GSoC2009.md b/www.suckless.org/GSoC2009.md
@@ -142,25 +142,41 @@ field.
We'd like to see a lightweight and simple alternative to gnome-volume-manager
and similiar programs.
-### A modern C99-only compiler
+### Yet another less sucking editor
+Although vi(m) does its job, it has become a monster during the time. We
+believe there is a gap between ed, sam, acme and vim which must be filled with
+a completely new less sucking editor.
-We'd like to see a modern C99 compiler which performs reasonable well and only
-supports C, not C++. gcc is too slow and too focused on language-agnostics and
-particularly focused on its C++ support. We got the impression that most open
-source software is written in C and makes no use of C++, so that it's
-desirable to have a modern C99 compiler for the x86 and arm architectures.
+### Less sucking C99 subset spec
+
+Although sufficient, C99 is not a perfect programming language. It suffers from
+legacy syntax and semantics and lacks features that make designing libraries
+and developing abstractions much easier. To address these shortcomings the
+student should survey the problems of C, evaluate possible solutions and
+enhance the language as a subset of C99.
+
+### Modern libc
+
+The standard libc is full of awkward and legacy concepts. We believe that
+recent approaches such as Google's bionic libc is a step into the right
+direction, though not radical enough. We think there is a great opportunity to
+implement a completely new libc which abstracts a very nice standard
+environment which can be implemented on modern OS platforms in a straight way.
+Our goal would be to achieve an initial proposal on Linux.
+
+### Improve tcc
+
+We'd like to see tcc being continued and improved. gcc is too slow and too
+focused on language-agnostics and particularly focused on its C++ support. We
+got the impression that most open source software is written in C and makes no
+use of C++, so that it's desirable to have an improved tcc.
We are also concerned about recent attempts to implement the C front-end of gcc
in C++. We believe that is a bad decision in general (due to demanding C++ as
bootstrapping environment) and would like to get rid from the gcc dependency
for these reasons.
-We'd like to see that the C99 compiler is able to build recent linux kernels as
-a test bed.
-
-Due to the high difficulty we believe that this project might consist of up to
-3 students concentrating on different parts (1 student for the front-end, 1
-student for the AST optimization/normalization, 1 or 2 students for the x86 and
-arm backends).
+We'd like to see that the improved tcc is able to build all suckless projects
+and perhaps the modern libc replacement.
[FAQ]: http://code.google.com/opensource/gsoc/2009/faqs.html#0_1_ideas_5167658354380897_772
