Hmm, I'll have to go back and see what I presented at EclipseCon. The funny thing is that the theme of the presentation was to show how to use the Eclipse CDT to develope Firefox source. It was supposed to reveal that you can use the CDT with large projects such as Firefox with little pain.
Building projects using make, especially with recursive makefiles such as you find in Mozilla, is indeed a slow process. We are working on an internal builder that will operate much like Visual Studio. However, I think you'll have a hard time convincing the Mozilla community to overhaul their build system. That is why we continue to support seemless integration with pre-existing build systems in the CDT.
The big feature with CDT 3.1 was a rewrite of the indexer framework. I am now able to index Firefox in about 10 minutes. You only do that once when you set up the project. After that you hardly notice the indexer running when you change files. The comment you mention is in regard to the old Full indexer that has been the bane of our existance for years. The new indexer should address these issues.
I don't know any technology that can debug Java and step into JNI code. I notice that the C# debuggers (.Net and Mono) can do that, but no-one has invested in a Java debugger that can do that. And, yes, if you know how to do it, contributing to the CDT project is the best way to get some fame.
Visual Studio continues to be the benchmark that we try to achieve with the CDT. We are also looking at the amazing features that the JDT provides. We've certainly come a long way in the four years we've been working on it. I've received alot of feedback that tells me the CDT is not a joke, and of course there are those who think it is. But such is life in the "open".
Slashdot Mirror
Hmm, I'll have to go back and see what I presented at EclipseCon. The funny thing is that the theme of the presentation was to show how to use the Eclipse CDT to develope Firefox source. It was supposed to reveal that you can use the CDT with large projects such as Firefox with little pain.
Building projects using make, especially with recursive makefiles such as you find in Mozilla, is indeed a slow process. We are working on an internal builder that will operate much like Visual Studio. However, I think you'll have a hard time convincing the Mozilla community to overhaul their build system. That is why we continue to support seemless integration with pre-existing build systems in the CDT.
The big feature with CDT 3.1 was a rewrite of the indexer framework. I am now able to index Firefox in about 10 minutes. You only do that once when you set up the project. After that you hardly notice the indexer running when you change files. The comment you mention is in regard to the old Full indexer that has been the bane of our existance for years. The new indexer should address these issues.
I don't know any technology that can debug Java and step into JNI code. I notice that the C# debuggers (.Net and Mono) can do that, but no-one has invested in a Java debugger that can do that. And, yes, if you know how to do it, contributing to the CDT project is the best way to get some fame.
Visual Studio continues to be the benchmark that we try to achieve with the CDT. We are also looking at the amazing features that the JDT provides. We've certainly come a long way in the four years we've been working on it. I've received alot of feedback that tells me the CDT is not a joke, and of course there are those who think it is. But such is life in the "open".