That is not what the man said. Some hints:Oh, ok, no help wanted, but it was a try
I asked for, but did not get an answer.Per wrote:Yes, patches should go to trac, not here.
Put them in one and paint on it => "Reduce MSVC warnings in general".Per wrote:They should also be split by type of change, and not by area of code they modify.
I am sorry folks, but i guess it is complete senseless to help you. It will take me longer to correct everything for each of you instead of just correcting these errors and warnings. I can swear, after i have split them again and erradicated the last three lines of "completely senseless style changes", the next one will jump out of the bush and help me to understand another tiny funny part which could be done better. It is seriously sad that you do not spend this energy in a better and a bit more stable hardened source code.Per wrote:These patches have warning fixes and code rewrites mixed up with completely senseless style changes like changing /* comments */ with // comments and re-indenting (not all of which follows CodingStyle).
Can you please resubmit the changes split by type of warnings fixed, and drop the cosmetic changes?
Sorry, you feel that way, but cleaning up the codebase by using a beautifier screws everything up. If you read the ML, I wanted to do the same thing, a long time ago, then realized that it just wasn't worth it because all of the issues that will cause.Seismo wrote:I am sorry folks, but i guess it is complete senseless to help you. It will take me longer to correct everything for each of you instead of just correcting these errors and warnings. I can swear, after i have split them again and erradicated the last three lines of "completely senseless style changes", the next one will jump out of the bush and help me to understand another tiny funny part which could be done better. It is seriously sad that you do not spend this energy in a better and a bit more stable hardened source code.
Well, except that from I look, we're all telling you the same thing, and as Buginator explained, it is for some very good reasons.Seismo wrote:I can swear, after i have split them again and erradicated the last three lines of "completely senseless style changes", the next one will jump out of the bush and help me to understand another tiny funny part which could be done better. It is seriously sad that you do not spend this energy in a better and a bit more stable hardened source code.
Ok we are running in a circle. If you can see the serious warnings instantly between the 230 reported you might be much better than me. By correcting the warnings i also found unicode errors (WCHAR to CHAR casts) in exeption handler lib and a beauty misconfiguration in netplay which conflicts with global settings. You might not be able to see that because you are using a different operating system and/or a different compiler than me. I always mentioned that this would be a help for the MSVC users (wich also implicit that they use at least windows as operating system). I also took care not to come into conflict with other compilers and systems by using the appropriate preprocessor globals, due to the fact that i can not guarantee for things i do not use or have.Per wrote:Well, except that from I look, we're all telling you the same thing, and as Buginator explained, it is for some very good reasons.
Also, adding casts all over the place does not result in a 'more stable hardened source code'. In fact, it is quite the opposite, as the code ends up more brittle when types are changed during development. The cast warnings should be turned off - implicit casts are allowed in C.
AFAIK, (nearly) all developers use gcc under Linux. There you have the option to turn certain sorts of warnings on and off via a compiler flag in the Makefile (as opposed to a pragma in the source code). Like lowering the warning level on MSVC, but more fine-grained. gcc with the current flags produces no warnings. And if warnings turn up, they're instantly corrected (standard cflags include -Werror - every warning counts as an error and immediately stops compiling).Seismo wrote:If you can see the serious warnings instantly between the 230 reported you might be much better than me.
