learning about the advantages of using a specific compiler are to my point of view not the first thing to know when switching to Linux (or to any OS). I would let that for later. There are so many other differences that will affect you before that.

Basically newer compilers should generate better machine code that runs better on given processors and have (in theory) less bugs. You should then get a little speedup using a new compiler in theory. Of course the results are different if your whole system is compiled with that compiler or not.

In terms of new Mozilla features, there shouldn't be any. The end-user features mozilla provide should (in theory) not depend on the type of compiler used. I didn't read all lines of code to ensure that. In terms of speed, see above. In terms of compatibility, that will depend what your underlying system is running. There have been some compatibilities changes in the ABI since 2.95.x and you will poobably run into trouble running an application compiled under 2.95.x on a 3.2.x system or the other way around. But these problems should be taken care by your distribution (Red Hat, Debian, ...).

If your system is using 3.2.x you can then try uswing those binaries. Red Hat 9 uses 3.2.1. If you have this one, get a go to the mozilla gcc323 binary.

Correct me if I am wrong.

I can't believe I am answering this as I don't know anything about it, but if you read the description of this bug http://bugzilla.mozilla.org/show_bug.cgi?id=204236, you'll see it had to do with licensing issues. This bug was one of the 1.4 blockers.

Just wait until they switch from -O to -O2 optimization http://bugzilla.mozilla.org/show_bug.cgi?id=158385 move to gcc 3.3 http://gcc.gnu.org/gcc-3.3/changes.html and optimize for Pentium Pro or Pentium 2 http://forums.mozillazine.org/viewtopic.php?t=13331&highlight=PPro Those changes should make Mozilla on Linux about as fast as on Windows.

When you compile with optimisations such as -O2, the executable's size grows. So you get bigger builds that actually run faster.

By the way: I've always found the size argument to be not so relevant. Everyone nowadays wants a full-featured browser that fits on half a floppy disk...

When you compile with optimisations such as -O2, the executable's size grows. So you get bigger builds that actually run faster.

By the way: I've always found the size argument to be not so relevant. Everyone nowadays wants a full-featured browser that fits on half a floppy disk...

Yep.

Redhat switched to gcc 3.x with Redhat 8.0, thus the change in major version number. Thus things compiled with gcc 3.x cannot be expected to run reliably on Redhat 7.x.