ping

On 11/24/17 13:13, Martijn van Duren wrote:
> On 11/24/17 12:07, Stuart Henderson wrote:
>> On 2017/11/24 09:27, Martijn van Duren wrote:
>>>> @conflict php-fastcgi->=5.6,<5.7
>>>
>>> Left out, since there's no overlapping names between the two packages,
>>> which is what pkg_create(1) specifies should be the reason of usage.
>>> Of course it's easy enough to add if you think it's still needed.
>>
>> Good point. As long as the upgrade test works that's alright.
>>
>>>>> +--- sapi/cgi/php-cgi.1.in.orig
>>>>> ++++ sapi/cgi/php-cgi.1.in
>>>>> +@@ -1 +1 @@
>>>>> +-.so man1/php.1
>>>>> ++.so man1/php-5.6.1
>>>>
>>>> Wrong version.
>>> Autch
>>> It was meant to read .so man1/php-${PV}.1. Probably got overwritten and
>>> overlooked during a final update-patches.
>>>>
>>>> Did you figure out why it was copied rather than using .so
>>>> before? There might be a reason for that.
>>>
>>> It wasn't installed at all before, so this is the first time php-cgi(1)
>>> is added. If it was copied I would've left it as is.
>>>
>>> On second thought I do now reckon it's better to just copy the original
>>> php.1. If -cli is going to be split of the .so file might not be there
>>> and the content of the file hasn't been changed since it's creation 5
>>> years ago so it's unlikely to get its own manpage any time soon.
>>
>> +1
>>
>>>     ${INSTALL_PROGRAM} ${WRKBUILD}/sapi/cli/php ${PREFIX}/bin/php-${PV}
>>> -   ${INSTALL_PROGRAM} ${WRKBUILD}/sapi/cgi/php-cgi 
>>> ${PREFIX}/bin/php-fastcgi-${PV}
>>> +   ${INSTALL_PROGRAM} ${WRKBUILD}/sapi/cgi/php-cgi 
>>> ${PREFIX}/bin/php-cgi-${PV}
>>> +# Make sure that php-cgi.1 still just sources php.1 when importing a new 
>>> major version.
>>> +   ${INSTALL_MAN} ${WRKSRC}/sapi/cli/php.1 
>>> ${PREFIX}/man/man1/php-cgi-${PV}.1
>>
>> It's not /all/ that likely that this comment will be noticed that far
>> down in Makefile.inc when someone adds a new branch. What I sometimes
>> do in cases where I'm worried upstream might change something that
>> might not be noticed and needs further attention is make an explicit
>> check that kills the build (e.g. see the pjproject check in Asterisk's
>> post-extract), might be worth considering. Then again it doesn't seem
>> all that important anyway..
>>
> I tend to sway to the latter. It's not really that important. I added this as
> a general reminder for future maintainers to check it occasionally and as a
> motivation why php.1 is installed instead of php-cgi.1.
> 
> So OK for the current version of the patch?
> 
> Also, any strong opinions about moving out -cli, -fpm, -apxs2 of -main and
> enabling -phpdbg in its own subpackage? I'll probably want to pick that up
> next.
> 

Reply via email to