Let's face it. The 32-bit Win32 builds have always produce 100 of warnings, even at their best, before they started consting everything in sight.
With the 64-bit builds it must be close to 2,000 warnings? I'll do realclean and count them later.
But can you imagine me submitting a patch that changed *every* source file? And 10s to 100s of lines in each source file.
I might just as well suggest putting the whole codebase through a C-beautifier to sort out the indentation :)
Update: 2106 warnings:
| [reply]
[d/l]
[select] |
Given that the perl community (in general) frowns so heavily upon warnings being emitted at *any* stage, it's rather surprising that this sort of thing is tolerated.
I never got the impression p5p is really concerned about bending code backwards just to avoid warnings. They are *warnings* after all. And there are (C) compilers that are picky. Very, very picky (they are C compilers, so they can spend time on nitpicking). Combine that with the fact the perl sources need to compile on a wide range of compilers, compiler versions, platforms and OSses, and it contains code that has been written eons ago. I'd very surprised if there are many compiler/OS/platform combinations that doesn't emit a single warning during the perl build process.
| [reply] |
I'd very surprised if there are many compiler/OS/platform combinations that doesn't emit a single warning during the perl build process
Yes, I get 3 compiler warnings on linux. However, I would think that if *nix builds generated as many warnings as the x64 windows build, p5p would do something about it quicksmart.
Personally, I'm not greatly perturbed - the main problem on x64 windows is that, because there's so many warnings, your chances of spotting the odd one out that *does* have a bearing on a problem is greatly reduced.
Cheers,
Rob
| [reply] |
However, I would think that if *nix builds generated as many warnings as the x64 windows build, p5p would do something about it quicksmart.
True, but that's something else than
Given that the perl community (in general) frowns so heavily upon warnings being emitted at *any* stage, it's rather surprising that this sort of thing is tolerated.
Personally, I'm not greatly perturbed - the main problem on x64 windows is that, because there's so many warnings, your chances of spotting the odd one out that *does* have a bearing on a problem is greatly reduced.
Yes. To bad that six release candidates (RC0 - RC5) were made available over a period of several weeks, and noone spoke up about this issue then. One of the reasons Perl .0 releases have a bad reputation is that release candidates are commonly ignored - and problems are only spotted afterwards.
| [reply] |