develooper Front page | perl.perl5.porters | Postings from March 2020

Re: Killing Symbian, DOS (DJGPP) and NetWare in 5.34?

Thread Previous | Thread Next
Sawyer X
March 21, 2020 13:15
Re: Killing Symbian, DOS (DJGPP) and NetWare in 5.34?
Message ID:

On 3/21/20 2:54 PM, Karl Williamson wrote:
> On 3/21/20 6:51 AM, Karl Williamson wrote:
>> On 3/21/20 6:38 AM, Sawyer X wrote:
>>> On 3/20/20 7:14 PM, H.Merijn Brand wrote:
>>>> On Fri, 20 Mar 2020 09:28:03 -0600, Karl Williamson
>>>> <> wrote:
>>>>> On 3/19/20 3:50 PM, Tomasz Konojacki wrote:
>>>>>> On Thu, 19 Mar 2020 20:15:54 +0000
>>>>>> (Dagfinn Ilmari Mannsåker) wrote:
>>>>>>> Hi Porters,
>>>>>>> We still have (no doubt bit-rotten) code to support Symbian, DOS
>>>>>>> and NetWare, but nobody has been submitting smoke or bug reports
>>>>>>> in years ( does not have them in its
>>>>>>> OS dropdown).
>>>>>>> I propose we kill these in 5.34, which means we should announce
>>>>>>> this in perl5320delta, like we did with a bunch in perl5160delta.
>>>>>>> We actually threatened to kill djgpp back then, but never did.
>>>>>>> Any objections?
>>>>>> NetWare and Symbian definitely should go because they're dead
>>>>>> platforms that have been unmaintained for years.
>>>>>> DOS (FreeDOS) and DJGPP are technically still alive but it indeed
>>>>>> seems they have no users, so I'm not opposed to removing it.
>>>>> Further research from Tomasz indicates, I believe, that if we got rid
>>>>> of symbian we could get rid of PERL_GLOBAL_STRUCT.  It was added
>>>>> mainly for mingw, which no longer needs it.
>>>>> I think getting rid of PERL_GLOBAL_STRUCT shows that there is real
>>>>> benefit to removing Symbian at least; that it does cause unnecessary
>>>>> work.
>>>> Now, *THERE* I hear a good argument for dropping Symbian.
>>> This is the most compelling argument for me too.
>> Note that I did not claim for sure that the only reason for 
>> PERL_GLOBAL_STRUCT is Symbian; only that it appears so based on 
>> Tomasz' doing some research.  I emailed Jarkko to make sure he 
>> doesn't know another, and maybe Dave Mitchell also has information 
>> about this.
> In the meantime, I got the reply below from Jarkko:
> Kill it.
> The only qualm I have is that the PERL_GLOBAL_* are a useful regimen
> of encapsulation: they allow completely removing global symbols of
> (lib)perl so that only a few entry point functions remain. Even without
> platforms *requiring* that, cleanliness like that is useful.

How often is it used in this manner, rather than to support those platforms?

Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About