Hi, No don't have time to bisect this also I'm no trained to do it I'm also fed up being told it's so easy cos its not and why don't you have one guy trained who can be contacted to do it for others? or show others how to do it? I'm the maintainer for Rome total war latest report is the install fails got to be a regression fix it worked on 1.44 Regards, Keith
On Thu, May 27, 2010 at 1:24 PM, Keith Muir <keith_kw_muir(a)blueyonder.co.uk> wrote:
Hi,
No don't have time to bisect this also I'm no trained to do it I'm also fed up being told it's so easy cos its not and why don't you have one guy trained who can be contacted to do it for others? or show others how to do it? I'm the maintainer for Rome total war latest report is the install fails got to be a regression fix it worked on 1.44
Wine-devel is for development discussion, not end user help. Regression testing is relatively easy, lots of users do it all the time. If you have trouble, consult the forum / wine-users for help. Keep in mind wine is volunteer project, so you can't expect people to do things for you, you need to help us help you. For starters, you could check if the demo has the same problem: http://www.gamespot.com/pc/strategy/rometotalwar/download_6105478.html?sid=6... that way someone else could run the regression test, since not everyone has the game. Though with a poor attitude like that, I'm not sure how many people would want to help you. Remember, you'll catch a lot more flies with honey than with vinegar. -- -Austin
Please ignore the mail from webmaster(a)dawncrow.de, i used the wrong account. Thx -- Best Regards, André Hentschel
Keith Muir wrote:
Hi,
No don't have time to bisect this also I'm no trained to do it I'm also fed up being told it's so easy cos its not and why don't you have one guy trained who can be contacted to do it for others? or show others how to do it? I'm the maintainer for Rome total war latest report is the install fails got to be a regression fix it worked on 1.44
If you are the maintainer, can you at least update the App Db entry to advise there is a regression and file a bug report (if you have not done so) and please, please, stop whining... Regression tests are not really that hard. I'll give you an example. I have been working on the EM_FORMATRANGE function and let it go for about six months due to other things going on in my life (something called El Tour de Tucson, you can Google it.) When I picked up the code, the Print Preview function did not work. I had a lot of code to bisect. I finished this in less than six hours at about 20 minutes a build.... However, if you cannot do this and if there is a demo available, advise if the problem/error exists with it. In my case, I was using a free program and its location is well known. James McKenzie
Regression tests are not really that hard. I'll give you an example. I have been working on the EM_FORMATRANGE function and let it go for about six months due to other things going on in my life (something called El Tour de Tucson, you can Google it.) When I picked up the code, the Print Preview function did not work. I had a lot of code to bisect. I finished this in less than six hours at about 20 minutes a build....
Everything is in the git bisect man page. (type "man git-bisect" in a console and read). There is nothing much to it - it just does a binary search through revision but do the search under your guidance: you answer yes or no whenever a build finishes and tests if that build "work" (as in for a specific problem) or not. On modern linux systems, ccache caches compilation results and git by default when switches versions update all files which changes to have the latest file stamps, so either way one does actually do a full build (but either read from the compiler cache or rely on make - and the makefile - being correct to have the right dependencies and only rebuild files changed).
participants (5)
-
André Hentschel -
Austin English -
Hin-Tak Leung -
James McKenzie -
Keith Muir