Supporting 'port livecheck installed'

[Blair Zajac]

> On Aug 31, 2019, at 11:15 AM, Chris Jones <jonesc at hep.phy.cam.ac.uk> wrote:
> 
> 
> 
>> On 31 Aug 2019, at 7:12 pm, Blair Zajac <blair at orcaware.com> wrote:
>> 
>> Hi,
>> 
>> Ok, I get that.
>> 
>> Can we come up with a solution that works for you and for people that want to see all outdated ports they have installed easily?
>> 
>> How about doing this?
>> 
>> $ port live maintainer:cjones | uniq -f 1
> 
> Ugg. Not keen.
> 
> I do understand where you are coming from, but the approach used here is common in all the py ports... I think if you want to improve this you should come up with something better that works as part if the python PG or similar. The above is just an annoying hack. 

Fixing the portgroup would be the best way. Something like having it dedup the requests will be best or report the livecheck against the main port.

In any case, I don’t have the cycles to work on this, so we’ll leave things alone.

Blair


> Chris
> 
>> 
>> Would that work for you? It prints only the first main port reports.
>> 
>> Blair
>> 
>>> On Aug 31, 2019, at 11:04 AM, Chris Jones <jonesc at hep.phy.cam.ac.uk> wrote:
>>> 
>>> Hi,
>>> 
>>> I use
>>> 
>>> port live maintainer:cjones
>>> 
>>> For that. For the py ports I maintain I do not want the above to give multiple reports for each version supported. Hence the livecheck none for these, so only the main stub port reports back.
>>> 
>>> Chris
>>> 
>>>> On 31 Aug 2019, at 7:02 pm, Blair Zajac <blair at orcaware.com> wrote:
>>>> 
>>>> Yes, however, it’s not as convenient as seeing what ports you have that have new upstream versions. I would rather take more computing time and resources then my time manually doing `port livecheck py-virtualenv` and the other ones. My time is much more valuable than 4 times the number of HTTP requests.
>>>> 
>>>> Blair
>>>> 
>>>>> On Aug 31, 2019, at 10:57 AM, Chris Jones <jonesc at hep.phy.cam.ac.uk> wrote:
>>>>> 
>>>>> It is intentional that the version specific py ports have livecheck none. The idea is you don’t need every one of the versions performing the check, only the non versioned port.
>>>>> 
>>>>>> On 31 Aug 2019, at 5:50 pm, Blair Zajac <blair at orcaware.com> wrote:
>>>>>> 
>>>>>> I care about such ports, e.g. py37-tensorflow and py37-virtualenv, so I would have that in requested. So that appears to not directly solve the issue.
>>>>>> 
>>>>>>> On Aug 31, 2019, at 9:16 AM, Jeremy Lavergne <jeremy at lavergne.me> wrote:
>>>>>>> 
>>>>>>> As an alternative or workaround, would narrowing to `port livecheck requested` be useful for you?
>>>>>>> 
>>>>>>> 
>>>>>>>> On 8/31/19 11:58 AM, Blair Zajac wrote:
>>>>>>>> Hi all,
>>>>>>>> It appears a number of ports, particularly py-* ports, use “livecheck.type none”. I find running ‘port livecheck installed’ a handy way to see if any of the ports I care about have an update, however, this doesn’t work for the py-* ones.
>>>>>>>> Can something be changed to support this? Would the livecheck code need to specially handle the port name when ${name} is a py??-* variant?
>>>>>>>> Blair
>>>>>>> 
>>>>>> 
>>>>> 
>>>>> 
>>>> 
>>> 
>>> 
>> 
> 
>