Front page | perl.perl5.porters |
Postings from December 2017
On 12/25/2017 10:29 AM, Zefram wrote: > Sawyer X wrote: >> perhaps we could provide temporary warning for now? > You mean... a deprecation warning? Yes. > Technically, if we revert perl_parse() to return zero for exit(0), > it is feasible for this to detect that perl_run() would do something > if called, and thus to warn that this bogus zero return matters. > Thus using an early exit(0) with the intent that INIT blocks or the main > program subsequently run could be deprecated. It's not feasible to do > the same for perl_run() bogusly returning zero, or for perl_parse() > bogusly returning zero when the subsequent use the caller makes of > the interpreter would be something other than just calling perl_run(). > But we haven't seen any breakage of those kinds (yet). For the purposes > of pure Perl code such as Module::Install, as opposed to arbitrary C > level embeddings, the feasible deprecation is sufficient. Wherever we can warn here instead of die, it is better. Where we cannot, we already have a large enough list of breakages to churn through. We could try again later when the list has been reduced.Thread Previous | Thread Next