Aaron Arvey wrote:
This is the third or fourth resend... anything I'm unaware of?
We're using gcov and lcov to measure how well the Wine test suite covers the wine source tree. Here's our first cut at making it easy to run wine compiled for coverage testing, and view the results using the simple text interface of gcov. A seperate patch for lcov will be released soon.
I think the issue with this patch is that it isn't it doesn't appear to be useful for a significant number of people and could be maintained as an external patch by the people who do think it's useful. Maybe you could try to persuade us as to why everyone should be using gcov?
Rob
On Tue, 01 Feb 2005 10:39:00 -0600, Robert Shearman wrote:
I think the issue with this patch is that it isn't it doesn't appear to be useful for a significant number of people and could be maintained as an external patch by the people who do think it's useful. Maybe you could try to persuade us as to why everyone should be using gcov?
I'd really like to see this one go in: I think it would not be hard to set up automated dumps of the gcov output and convert them to something nice and easy to read. For people who want to volunteer but aren't sure where to start, or who aren't quite ready to dive into app compat work extending the test suite is a great thing to do.
In particular it'll let us easily identify things like error codes that may be incorrect (ie, were guessed). Given the number of times we've been nailed by people guessing error codes for obscure situations and not marking them as guesses, any tool that can show what we're not testing gets a +1 in my book.
Also the patch looks small, it's mostly documentation, and minimally invasive. I see no reason to reject it, hopefully Alexandre has just been bogged down with the WM rewrite and the huge volume of patches going in lately.
thanks -mike
Mike Hearn mh@codeweavers.com writes:
Also the patch looks small, it's mostly documentation, and minimally invasive. I see no reason to reject it, hopefully Alexandre has just been bogged down with the WM rewrite and the huge volume of patches going in lately.
The changes may be small but they are ugly IMO. I don't think we need a configure option for that, it should be possible to simply do a 'make CFLAGS=xxx' with the appropriate options.
On Wed, 2005-02-02 at 19:32 +0100, Alexandre Julliard wrote:
The changes may be small but they are ugly IMO. I don't think we need a configure option for that, it should be possible to simply do a 'make CFLAGS=xxx' with the appropriate options.
OK, I don't really care how it's activated as long as it's documented :)
I'm not sure how to avoid the problem with the preloader without build system changes though. Maybe just sed them out in loader/Makefile.in?
thanks -mike
-
Alexandre Julliard -
Mike Hearn -
Robert Shearman