On Monday 15 October 2007 15:41:07 Andrew Hood wrote: > Ivan Skytte Jørgensen wrote: > > I gave the I also gave it a spin with Flexelint 8.0w. It complains a lot, > > and I have included the most interesting warnings. [snip] > > The interesting warnings are: > > ==snip== > > I went through the entire source over the weekend with CFLAGS containing > "-Wuninitialized -W -Wall -Werror" ( I normally only use -Wuninitialized ). > > At a rough guess there are a couple of thousand lines that require > fixing. I can't CVS diff the source because it wasn't a clean checkout. > > Does this qualify as an "interesting thing"?
Probably. What did you find of genuine errors? The problem with aggressive warning generation is the number of false positives in mature code. On a related note, GCC has improved its warning generation a lot during the recent years. It still lacks laser-precision warning suppression, but it is a vast improvement over, say, version 2.95.3, which I am forced to use on a daily basis. The HP compiler ACC B.06.x series isn't bad either. It is starting to show value tracking, but so far it has not found a genuine error (yet). ACCs warning suppression is essentially non-existant. Regards, Ivan -- NEdit Develop mailing list - [email protected] http://www.nedit.org/mailman/listinfo/develop
