On 30 January 2012 14:33, Rafael Garcia-Suarez <rgs@consttype.org> wrote: > On 30 January 2012 14:22, demerphq <demerphq@gmail.com> wrote: >>> IMO, depraction warnings belong in linters. "use warnings" isn't a linter. >>> I think that, for those that have warnings enabled, the emitting of warnings >>> on each and every run of the program will be considered as bad as a breakage. >> >> I just hope that people agree that it warning once per line using a >> deprecated feature, and not once per invocation of that feature is the >> lessor of two evils. > > I think most deprecation warnings are compile-time, so they'll warn > once already. Ah, so making ALL of them like that would not be a big change! And would make their behavior consistent! So there really isn't much ground to object to it. :-) > This is not the case for at least : > - "goto" to jump into a construct (runtime warning) > - Use of *glob{FILEHANDLE} is deprecated (runtime warning) > - chdir('') or chdir(undef) as chdir() (runtime warning, should be > easy enough to move into a ck function) > - Use of inherited AUTOLOAD for non-method %s::%s() is deprecated > (runtime warning) Well the one I specifically had in mind was the regex one for unescaped curly-braces. Afaiui those would warn every time too. And I bet people would appreciate if those warning stopped happening so much. :-) cheers, Yves -- perl -Mre=debug -e "/just|another|perl|hacker/"Thread Previous | Thread Next