Dev meeting 2016 04 26

Agenda

• Release schedule https://github.com/bolt/bolt/issues/5239
• 3.0.0 release
  • Volunteers to update documentation in first couple of weeks after 3.0.0 release
  • Bug team
  • Blockers:
    • #4095 Thumbnails/Uploads/FileManager with remote filesystems
    • #5199 [Upgrade] Fatal error in dashboard from getLatest()
    • #5242 More v2 extensions showing up on v3 extend page

Actionable Items

Log

<gawainlynch> ping Bopp (who won't be attending) carsonfull gawainlynch phillipp rarila (if you have a pinger on) rixbeck rossriley SahAssar slick0 
<phillipp> pong
<SahAssar> pong
<gawainlynch> …and then there were 3
<phillipp> yeah, i just looked if we have a dev meeting today :D
<phillipp> so quiet everywhere
<gawainlynch> Yeah… King's Day here
<gawainlynch> OK… we might have the shortest one yet
<rossriley> checking in
<rossriley> God Save the King
<SahAssar> Well, we had a server blow up again at work, so it was loud at work at least.
<gawainlynch> Wait… really loud?
<gawainlynch> As in literally 
<phillipp> she shortest one was about 9 mins, right?
<SahAssar> nah, figuratively blow up.
<phillipp> *the
<gawainlynch> OK
<SahAssar> Anyway, lets get on it
<phillipp> yeah, lets get started
<gawainlynch> OK… SahAssar issue side
<gawainlynch> Also, everyone (who is here)… any objections to #5239
-[BoltIssueBall]/#boltcms- #5239 [open] [RFC] Tweaking the Release Schedule  https://github.com/bolt/bolt/issues/5239 
<SahAssar> well, the blockers are down to quite a few... I'm planning to do a real cleanup after 3.0 is out and close a lot of the old ones
<carsonfull> here
<gawainlynch> FYI… I just got a … there you are
<SahAssar> 5239 sounds good to me
<slick0> ohai
<gawainlynch>  /kick slick0 
<gawainlynch> <no blade>
<slick0> /kick gawainlynch YOU SAID THE WORD!
<gawainlynch> SahAssar: Yeah… just not convinced of a May 1/2
<phillipp> *is quiet*
<gawainlynch> carsonfull: Now you are here, and short on time, #4095 
-[BoltIssueBall]/#boltcms- #4095 [open] [Feature] Thumbnails/Uploads/FileManager with remote filesystems https://github.com/bolt/bolt/issues/4095  — assigned to CarsonF
<carsonfull> Haven't touched it
<carsonfull> Been swapped with GMO server stuff
<rossriley> gawainlynch: thumbs up from me for 5239
<gawainlynch> rossriley: Thanks mate!
<gawainlynch> OK… so let's get onto that
<carsonfull> I'm doing my best, but I just haven't had time lately
<gawainlynch> The constant stream of "Gawain" comments should sum it up
<carsonfull> Unless you guys want to start paying me :P
<gawainlynch> But I just want to get us to a short-sharp turn around
<rossriley> is there anything in 4095 that is blocking though? 
<gawainlynch> Nah… just PHPDoc to cover off BC
<gawainlynch> I just wanted to give Carson a voice
<carsonfull> PHPdoc? I tihnk it's more than that
<rossriley> ok, so we can just get it added as it’s done then...
<carsonfull> We just need to make sure the code is updated for new filesystem
<carsonfull> ie try/catch instead of bool returns
<gawainlynch> carsonfull: I thought you were going to …
<carsonfull> Yeah it's on my list if that's what you mean? 
<gawainlynch> No, mixe understanding of where you're at
<gawainlynch> *mixed
<carsonfull> Well I'm not sure if all the core code has been updated
<gawainlynch> Last PM session was "done but what about BC" (my words)
<carsonfull> For FS
<gawainlynch> OK… so Bopp is aiming to release in 6 days
<carsonfull> I should have time tomorrow to look at it
<gawainlynch> I am aiming to have a coronary about that point
<carsonfull> Frank only just merged my S3 change a few days ago
<carsonfull> I haven't been able to test core until that point
<gawainlynch> Yeah, I know… with you and in support
<gawainlynch> We just need to define a way forward
<gawainlynch> If it blocks, it blocks
<gawainlynch> But we need to reasonably define that
<carsonfull> Let me play with S3 in core tomorrow and I'll document on that ticket what I find
<carsonfull> Is that ok?
<gawainlynch> You have my complete support
<gawainlynch> Thing being that there are stakeholders being stakeholders
<gawainlynch> Let's just give them "something"
<gawainlynch> i.e. reasonable answers
<gawainlynch> rossriley: #5199 
-[BoltIssueBall]/#boltcms- #5199 [open] [Upgrade] Fatal error in dashboard from getLatest() https://github.com/bolt/bolt/issues/5199 
<gawainlynch> Happy to do the work
<gawainlynch> Just need to get clear on "what"
<carsonfull> Sure, let's talk offline about how you can help
<gawainlynch> carsonfull: Deal!
<rossriley> gawainlynch: what’s the cause of that error, is it just not running a dbupdate after doing an upgrade?
<gawainlynch> Yes
<gawainlynch> But if you update and then try to log in… *boom*
<gawainlynch> i.e. nut and you're OK
<rossriley> ah ok..
<gawainlynch> without nut, you're going to another platform, IRC, or Slack
<gawainlynch> I think we're more or less agreed that exceptions are good
<gawainlynch> But there needs to be transition paths
<gawainlynch> I feel really dirty when I add `catch () {}`
<gawainlynch> (often)
<rossriley> I guess, but in reality that’s what needs to happen… unless we are going to auto run the dbupdate
<gawainlynch> Yeah, but for everyone upgrading from 2.x to 3.x, we then require a nut
<gawainlynch> …to maintain the Bolt :-P
<rossriley> could the catch redirect them to the /dbupdate page
<carsonfull> I would try/catch and return []
<carsonfull> Maybe with a flash message of what ross said ^
<gawainlynch> OK… WFM
<gawainlynch> Lat one
<gawainlynch> *last
<gawainlynch> #5242
-[BoltIssueBall]/#boltcms- #5242 [open] More v2 extensions showing up on v3 extend page https://github.com/bolt/bolt/issues/5242 
<gawainlynch> rossriley & phillipp ?
<rossriley> that one is kind of how it’s designed at the moment..
<rossriley> the search box just reads this url: http://extensions.bolt.cm/list.json
<rossriley> and then the browse versions will show you compatible versions
<gawainlynch> OK… without putting in *lots* of work… can we fix it?
<gawainlynch> "fix"
<carsonfull> Is list.json custom or statis?
<rossriley> i don’t think so...
<rossriley> because the compatibilty search is more complicated
<rossriley> and when the package is searched for it only needs to happen on a single package...
<rossriley> whereas if we add it to the first box then we’d have to do a compatibility search on every single package
<gawainlynch> But at that point don't we know constraints?
<gawainlynch> Oh
<gawainlynch> Composer/Satis doesn't give us that then?
<carsonfull> What about /v3/list.json? Which is pre filtered 
<rossriley> yes, I guess we could do something like that where we have a separate list of v3 packages
<gawainlynch> We'd need to do magic
<carsonfull> That way we could parse in php with composer's semver code
<gawainlynch> Yeah… it heads into feature territory though
<carsonfull> More of unforseen bug I feel
<carsonfull> It doesn't provide anything more for the user
<gawainlynch> No me who would argue this down
<gawainlynch> I don't think we're quite ready for release
* carsonfull nods
<gawainlynch> OK
<gawainlynch> Actually
<gawainlynch> Stepping back
<rossriley> i don’t see a massive problem with how it is now to be honest.
<carsonfull> What happens with the v2's when you try to install them?
<gawainlynch> Well to be fair I've tripped over it 
<gawainlynch> carsonfull: Bolt dies hard
<carsonfull> Can you click install and then composer errors?
<rossriley> as soon as you browse versions for the package it will tell you which ones are compatible
<SahAssar> Well, it does not show the versions since none are compatible?
<rossriley> no, you can't
<gawainlynch> I had v2 extensions install today… on master & release/3.0
<carsonfull> Wait composer actually installs it??
<gawainlynch> Mmmhmm
<rossriley> no, you can’t install?
<SahAssar> gawainlynch: proper releases or dev-masters?
<carsonfull> Why doesn't that fail bolt/bolt version constraint?
<rossriley> The versions listed below are marked as stable releases by the developer and are compatible with the version of Bolt you are currently using….
<rossriley> No Stable Versions Available
<rossriley> that’s what you see when you try to install a v2 extension on v3
<gawainlynch> dev-master with requires of ^2.0
<carsonfull> dev-master for core?
<SahAssar> I'm getting "No Stable Versions Available" and "No Test Versions Available"
<rossriley> ^ 
<gawainlynch> master & release/3.0
<phillipp> when you check for available updates of already installed packages, you will get bolt 2 versions
<gawainlynch> OK… happy to go with PEBCAK
<gawainlynch> Yeah, that too, Phillipp
<phillipp> on bolt 3 beta 2 it suggested me to update to an older version of gawains members extensions
<phillipp> i think there is also an issue from me
<phillipp> #5184
-[BoltIssueBall]/#boltcms- #5184 [open] [BUG] Check updates shows wrong new version for bolt/members https://github.com/bolt/bolt/issues/5184 
<gawainlynch> Yeah… that is kinda rossriley point
<gawainlynch> (insert missing plural)
<gawainlynch> OK… I don't think there is the momentum to grab this one
<gawainlynch> May 2 release…
<gawainlynch> Final thoughts?
<phillipp> i have no problem when it takes a bit longer
<gawainlynch> Yeah… others do and we need to argue well
* gawainlynch thinks we're not there
<SahAssar> I have no problem pushing as long as we set a firm date to push to.
<rossriley> I’m in favour of going ahead with the release...
<carsonfull> Idk why we would spend a year on this and then blowing it at 98%
<rossriley> we do need a warning for people to check extension compatibility before they upgrade though
<gawainlynch> carsonfull: Solutions though
<carsonfull> Fix the bugs first
<gawainlynch> rossriley: Yeah… kinda the last conversation for me… just not sure how to go about it
<gawainlynch> *kinda the point of the last…
<SahAssar> When it comes to the discussion about fixing stuff before release... #4503 is looking pretty bad
-[BoltIssueBall]/#boltcms- #4503 [open] [Tracker] Documentation updates https://github.com/bolt/bolt/issues/4503 
<SahAssar> Although, some things have been fixed, just not check off the list
<gawainlynch> OK… so without a way to negate this reasonably and constructively, the release manager is going ahead on May 2
<gawainlynch> #karma SahAssar 
<[BoltIssueBall]> BoltKarma for SahAssar is now 134
<carsonfull> #karma SahAssar
<[BoltIssueBall]> BoltKarma for SahAssar is now 135
<gawainlynch> OK… so we have a not-insignificant amount of the core team in favour of release… if we're going to modify that we need a proactive solution based reason
<carsonfull> Who's for?
<carsonfull> Or how many out of total
<SahAssar> 2 / 7, right?
<carsonfull> ^
* SahAssar is bad at maths though :P
<gawainlynch> OK, well Bopp & rossriley are "for"… so far against (of those speaking up) is phillipp, carsonfull & SahAssar (I abstain for balance), rarila and slick0 are MIA
<gawainlynch> So that is a 40/60 spilt with a deciding vote in my hands
<carsonfull> Not really. If you were for it would be 50/50
<gawainlynch> But lets not get that petty 
<gawainlynch> Yeah, my point
<SahAssar> Alright, basically the situation is that it does not seem like any side is budging on this. If the date is unmovable we need to deal with the deadline, not the reason for it.
<gawainlynch> Give me the space to let y'all decide and I'll break
<carsonfull> I think we are going to be creating a lot of extra work for ourselves answering questions if the docs aren't done
<gawainlynch> carsonfull: I agree btw
<gawainlynch> (tie-break btw)
<gawainlynch> <insert missing comma)
* gawainlynch needs to learn to type
<gawainlynch> OK… so… if we're going to block… solutions
<gawainlynch> i.e. let's not be Drupal, but if we're goin to block, let's come up with something workable 
<rossriley> what would be helpful is if we could realistically tag things that are genuine blockers...
<gawainlynch> +1
<rossriley> for instance in the documentation, a load of stuff could be covered with the links to the demo extensions...
<rossriley> I’m not convinced that there’s that many holes
* btmash (~btmash@drupal.org/user/60422/view) has joined #boltcms
<carsonfull> Everywhere I look in the docs there is old outdated info
<gawainlynch> rossriley: I tend to agree on that point
<rossriley> and I’ve got one eye on the fact that we won’t really get any proper wide production testing until we open it up to  release
<gawainlynch> One thing I have lacked is feedback on where people *might* struggle
<SahAssar> For example, has routing changed that much that it needs new docs for example?
* SahAssar retypes "for example" again for example
<gawainlynch> rossriley: +1
<gawainlynch> OK. so
<SahAssar> So, weed out what does not need docs or at least does not urgently need them on release
<SahAssar> Fix #5199 so we can have composerless upgrades
-[BoltIssueBall]/#boltcms- #5199 [open] [Upgrade] Fatal error in dashboard from getLatest() https://github.com/bolt/bolt/issues/5199 
<SahAssar> *nutless
<gawainlynch> OK… but how are we going to define what is a release target?
<carsonfull> gawainlynch: What do you mean?
<gawainlynch> We can't just "object"… we need to get this out the dor… 98 or 100%
* gawainlynch needs to learn to type
<gawainlynch> carsonfull: Just that let us all object *and* successfully block
<gawainlynch> But to do that, we need to come up with a release plan
<gawainlynch> …and make it happen
<carsonfull> You mean a date?
<gawainlynch> "plan" … "strategy" type things
<gawainlynch> …skip the "wank"… just how do we get over the line
<gawainlynch> We can't delay forever
<gawainlynch> We are 14/15 months into this branch
<carsonfull> That doesn't matter. If it's buggy or incomplete it's not ready for release
<gawainlynch> There are 3 blocking issues that could close soon (next few days)
<gawainlynch> Yeah, it does… we need to either release or come up with a solid plan to do it
<gawainlynch> If there are enough reasons to now… OK, don't… but how do we get this out?
<gawainlynch> i.e. let's combine heads here
<carsonfull> Why isn't the list of blocking issues enough?
<SahAssar> We decide what docs are absolutely *required* and get those in. We get the blockers that are actually blocking fixed.
<gawainlynch> Currently they can be easily closed off
<carsonfull> If someone wants to argue whether the issue is actually blocking or not let's do that on the issue
<gawainlynch> …and that one that can't we discussed as "isn't"
* ManTK is now known as ManTK`zZzZz
<gawainlynch> carsonfull: You should have a look at Drupal 8 ;-)
<gawainlynch> We're tired… we're worn out… none of us need more crap… we *all* agree
<gawainlynch> Honestly, I want to crawl in a hole for 2 weeksn
<rossriley> ok, so I’d argue that #5242, #5161 and #4095 aren’t blockers
-[BoltIssueBall]/#boltcms- #5242 [open] More v2 extensions showing up on v3 extend page https://github.com/bolt/bolt/issues/5242 
-[BoltIssueBall]/#boltcms- #5161 [open] Create Project from single composer.json throws errors https://github.com/bolt/bolt/issues/5161 
-[BoltIssueBall]/#boltcms- #4095 [open] [Feature] Thumbnails/Uploads/FileManager with remote filesystems https://github.com/bolt/bolt/issues/4095  — assigned to CarsonF
<SahAssar> Agreed
<rossriley> #5117 is easily fixed, as is #5199
-[BoltIssueBall]/#boltcms- #5117 [open] Database entities won't hydrate if columns use captials https://github.com/bolt/bolt/issues/5117  — assigned to rossriley
<gawainlynch> carsonfull: ?
-[BoltIssueBall]/#boltcms- #5199 [open] [Upgrade] Fatal error in dashboard from getLatest() https://github.com/bolt/bolt/issues/5199 
<gawainlynch> 5199 is the one that worries me
<gawainlynch> 5242 as well, but I can live with it
<carsonfull> I'm fine with not calling 4095 a blocker. Just know that that's my first priority
<gawainlynch> carsonfull: We do, and I know it will be done very soon
<gawainlynch> carsonfull: Honestly, that one is so "done" it isn't funny… you're all over it
<gawainlynch> That said… Bob is still release manager
<carsonfull> Yeah. Now that S3 is updated, it's almost there
<carsonfull> Where is Bob?
<gawainlynch> It is King's Night here
<gawainlynch> Only the biggest single days of the year in NL
<SahAssar> Right, I'm unblocking those three.
<gawainlynch> Tonight is 6 million Dutch people vomiting in the street :-D
* gawainlynch didn't say that
<gawainlynch> OK… we've hit 60 mintus
<gawainlynch> *minutes
<gawainlynch> Let's pull this in
<gawainlynch> SahAssar: +1
<gawainlynch> Support is here… just need a way to the line
<gawainlynch> (as in, I don't have one myself)
<carsonfull> How to define ready? When there are no more blocking issues
<gawainlynch> carsonfull: Which are being closed as we speak
<gawainlynch> (as in the open ones)
<carsonfull> Cool
<gawainlynch> Mate… if we're going to block… we need reason, and solution 
<carsonfull> Block what?
<gawainlynch> Release
<gawainlynch> Or are we all good now?
<carsonfull> For release?
<gawainlynch> Yes
<gawainlynch> (corret)
<gawainlynch> (correct)
<carsonfull> No there are still open block release issues
<gawainlynch> Which have been flagged in this meeting as easy fixes for this week
<carsonfull> Great once those are closed then we are ready
<gawainlynch> 5117 is so corner case it isn't funny anyway 
<gawainlynch> OK
<SahAssar> docs are now the biggest blocker IMO
<gawainlynch> SahAssar: OK… so what I said to carsonfull … solution?
<gawainlynch> (not "you have to do it")
<gawainlynch> i.e. what gets you across the line?
<SahAssar> So, what do we have in docs that is less documented than their equivalent in 2.2? IE not what do we want in there, what do we need in there? That's the question.
<gawainlynch> I don't think there is
<gawainlynch> Honestly, v3 docs IMHO are already better that v2
<SahAssar> I'm not as familiar with it since I haven't used 3.0 in production yet, so is there any of those items that need to get done before a release?
* gawainlynch doesn't read docs
* gawainlynch reads CakePHP code for inspiration 
<phillipp> lol
<gawainlynch> OK
<gawainlynch> So
* SahAssar is fine then
<gawainlynch> OK
<gawainlynch> So
* SahAssar is fine then
<gawainlynch> </meetin> #beer