Front page | perl.perl5.porters |
Postings from October 2012
On Sunday, 21 October 2012, Steffen Mueller wrote: > On 10/21/2012 05:21 PM, Vincent Pit wrote: > >> Hmm, point taken and slightly surprised by the vehemence of your >>> response. I respectfully disagree, though. >>> >> >> Sometimes I have definite opinions on things, and sometimes I voice >> them. They may or may not overlap with yours. >> > > Just for the record in case it wasn't clear: There wasn't supposed to be > even a hint of criticism in those two sentences I wrote. > > I just don't want Perl to support silent code breakage, especially when >> the only motivation is a slight performance improvement. >> > > Perl supports many, many, many ways to have silent code breakage. But > admittedly they're generally of different types: Within the language. > > Two compromises that would make sense to me: >>> >>> 1) Only support a configure flag for the variant where -t/-T is fatal. >>> Keep the Perl-internal, undocumented (but obviously named) define so >>> that whoever wants to use the silent variant doesn't have to maintain >>> patches against Perl or if anything, only patches that are limited to >>> very simple and localized changes. >>> >> >> What I objected against was it being *non fatal*, so I'm fine with this >> as long as it's not documented and not available through Configure. But >> note that this adds maintenance cost to the core for an undocumented >> feature, and I thought the trend was to go the other way. >> > > The code difference between the fatal and the silent variants is very > small. Not a high cost. > > 2) Support two variants in Configure: fatal and warning (no silent >>> variant). >>> >> >> At the time the warning is printed it's already too late, so that's as >> bad as no warning at all. And I guess that those who want the silent >> variant will not be satisfied by seeing that warning every time. >> > > It's a warning that would be useful in pre-production, not production. So > it's not too late. Furthermore, I think there's a class of cases where such > a warning is fine. It's not silent. If you run a script that actually makes > real use of -T, you'll know. But it doesn't prevent various test suites of > CPAN modules from running just because they author figured -T would be a > good idea in a test's shebang. > One thing that worries me is dangerous tests that are expected to be blocked by tainting being run just fine with the special perl build.... Not that I see it as a show stopper. Yves -- perl -Mre=debug -e "/just|another|perl|hacker/"Thread Previous