2019-08-01 - Re: [GRASE-Hotspot] Update - Have your say on future development

Header Data

From: Tim <ti***8@gmail.com>
Message Hash: 5dc7ea32bc066fbbe4f2dac392dadb93567a38b68b3596abdd7c8204dac92f74
Message ID: <CAESLx0+aVU85TxWoMKb8C4ZPLe_FN9QOzyaySa+BH3c9rvu=Aw@mail.gmail.com>
Reply To: <CAGPMUcEtr16vPRVV7kzkV5pEDYK=CJMACMLO14Qn-9kjSh33wg@mail.gmail.com>
UTC Datetime: 2019-08-01 23:42:00 UTC
Raw Date: Fri, 02 Aug 2019 16:42:00 +1000

Raw message 

On Fri, 2 Aug 2019 at 11:16, Charles Chambers <cc***2@gmail.com> wrote:

>
> 2)  The underlying dependencies (coova-chili and the rest) as required are
>>> also unmaintained versions dating from *prior to* 2014.  In other words,
>>> the underlying dependencies are now also abandoned.
>>>
>> I follow instructions literally, usually, and I remember briefly
> validating and verifying the versions.  There is enough of a mismatch
> between what you have customized and what is out there and what has not
> been updated in a few years to cause concern.
>
> Interesting. It does appear that Coova Chilli hasn't had a release since
>> 2016. However, development is still occurring in the repo, sporadically. I
>> build the Coova Chilli packages that Grase depends on, purely so we can
>> make sure the features we need are present. For each Grase release, I
>> attempt to build the latest Coova Chilli package.
>>
>
> See above.
>

At the point I can build an 18.04 package, I'll upgrade Coova Chilli to the
latest version. Each version changes things slightly, so trying to rebuild
Coova Chilli for a new image takes a considerable amount of time.



>
>> 3)  *Tim* is using a development framework dating from 2014 (or
>>> earlier).  As I understand it, he needed/needs to update that to go with a
>>> change implemented by Canonical in version 16.04 (three years ago).   We've
>>> been waiting for *that* update since 2016, three years ago.   It hasn't
>>> happened yet.
>>>
>> The original code had no framework, and relied on MySQL/PHP things that
>> aren't supported in PHP 7. 16.04 onwards ships with PHP 7, which is a good
>> thing, but makes the job of updating Grase to PHP 7 a pain. For that
>> reason, the update that I wanted 3 years ago essentially has turned into a
>> massive rewrite of core functionality on the Symfony framework. Work on
>> that is in a branch (
>> https://github.com/GraseHotspot/grase-www-portal/tree/symfony4)
>>
>
> My bust.  I'm *terrible* with names <g>.
>
> That IMHO is the most important piece to complete.  Even if the rest does
> not change, getting to a point that it works on at least 18.04 is simply
> the most important step in the process.
>

I did start down the track of doing the least work possible to run on PHP
7. Unfortunately for the database side of things we relied on a PEAR DB
library, which was a great idea back them before PDO existed, and was the
worst idea for trying to move to PHP 7. So by the time you rip out that
layer and replace it with something modern, you've replaced all the
important logic anyway. Then you discover that the templating engine also
doesn't play with PHP 7 nicely, so you rip that out and suddenly you have a
brand new application. Essentially, without trying to run a really old
version of PHP on 18.04 (I'm not going to try and package PHP 5 for 18.04),
we're stuck in the "rewrite" we're in. Any code that can be reused has been
reused, but all the database and template code has to be redone on a modern
library, so the Symfony framework is a freebie that comes with that.


>
>
>
>>
>> 4)  Everything is set up as open software, so developing a derivative
>>> package that bundles everything together is permitted, and the source code
>>> is also freely available.
>>>
>> Correct. More recent branches have gitlab CI files as well that should
>> help with building the packages. If anyone wanted to help, they don't need
>> to fork it, just talk to me and start helping. Forks are welcomed as well
>> though.
>>
>
> Ultimately it should finish up as the package, dependencies as drawn off
> of the Ubuntu repositories, and customization after that.  I haven't look
> at package architecture in quite some time, but IIRC it's possible to
> install the base package, draw and install the dependencies off the Ubuntu
> repository, and then draw custom configuration files out of your archive.
> I'll see what I can find.  Intricate, but done once.
>

Most of the dependencies we relied on previously didn't exist in Ubuntu, or
like Coova Chilli were old versions. Anytime I don't have to package a
dependency and can use Ubuntu's I will.



>
> Put simply. I'm time poor, and I need to pay the bills which dictates
>> where my time needs to be spent. I have no need for the Grase hotspot
>> anymore, in fact my niche requirement for it was removed 8+ years ago. I'm
>> happy to continue trying to plod along if the community wants it, but
>> unfortunately I need income to support my family if I'm doing that. If the
>> community wants me to abandon it, I will. I'm not going to try and give
>> projected update dates, because I've learnt that no matter how good I am at
>> estimating how long a project should take (that is part of my job after
>> all), I'm really not good at estimating how much free time I have to work
>> on it.
>>
>
>
> I have no particular problem with supporting the upgrade to 18.04, if the
> rest of the community can get behind it as well.  If I'm viewing the
> project correctly, you're looking at two months FTE to bring it current,
> and I can't see being the only one to support the entire upgrade.
>
> Bear in mind, I haven't deployed it yet, and as I work with it I get
> closer to deployment.  *THE* big holdup is a lack of self-registration.
>

What kind of self registration? There is already the "free account" button
that allows devices to get a "free" account based on a group (so
restrictions are applied). I also have some patches I've done over the
years that aren't good enough to make it into the release (hacky patches)
that would help you get started with other registration methods (like
needing to provide name/email first).


>
> Also, please don't take this as an attack on the community. The community
>> have done a wonderful job over the years including translating into many
>> languages, and really helping each other on the mailing list by answer
>> support questions, trying to work through strange hardware issues, etc. If
>> the community hadn't sprung up, I would have stopped development 8 years
>> ago.
>>
>
> I'll take my pique as impatience, and accept your response as chiding.
>

I want to just say Thankyou for asking the question. I'm not great at
giving updates until asked, so it's good to get the prompt to give an
update. It also gives me another chance to see what the community interest
is like. People asking for updates and still wanting to use it, that helps
me know that it's still wanted. No point developing something that's not
still wanted, especially when I see other solutions out there that are
"good enough" for some people.

Maybe I can work out an interim solution of PHP5 on 18.04, I'll have a play
around with it in the near future. But it'll be 'unsupported', best effort
kind of thing (especially as it'll involve installing a PPA to get it
working).

Regards

Tim

Thread