Bill, I think this sounds fine, but would make two possible suggestions.

1) I think we should get some kind of release built before bug-squashing week, 
to be usable for anyone wanting that sort of testing path during the following 
week.  That could easily happen late on the 7th, or we could XUL/Angular merge 
on the 6th and have the 7th to build.  Whether this is branded as an alpha or a 
beta1 doesn't matter too much, though if it is "feature complete", beta1 seems 
more correct.

2) It is reasonable to expect this might lead to a need for a beta2, but my 
preference would be to make that call a little later on, as needed.  It is easy 
to add more delays, but assuming we'll need them all now means we'll never get 
the time back :)

Just my two cents.

Dan

________________________________________
From: Open-ils-general <open-ils-general-boun...@list.georgialibraries.org> on 
behalf of Bill Erickson <beric...@gmail.com>
Sent: Tuesday, August 28, 2018 6:11:01 PM
To: Public Open-ILS tech discussion
Cc: open-ils-general@list.georgialibraries.org
Subject: [OPEN-ILS-GENERAL] 3.2 Release schedule extension (was: feature        
freeze and more)

Breaking this message out specifically to discuss extending the 3.2 release 
schedule.

We have a lot of competing priorities at the moment.  This week really should 
be about wrapping up feature merges, but the pending Beta is forcing us to 
address a number of outstanding issues, and each depends on the other.

Extending the release schedule seems perfectly reasonable to me.  My only 
concern is determining how best to leverage the Sept 10-14 bug squashing week.  
Ideally, all major changes would be merged so we can work out the kinks during 
the group bug squashing.

That means features, XUL removal, and Angular merging would need to be done by 
the end of next week, with time to spare to ensure all this chaos leaves us 
with a usable code base for testers.

To buy us some breathing room in the short term, I'll make this proposal:

Feature freeze pushed to Sept 4th.
XUL and Angular merge freeze Sept 7th.
Bug Squashing Sept 10-14
Beta Sept 19th (remaining targets pushed back 2 weeks as well).

This may not be the extension you were hoping for Kathy, and we can certainly 
modify this, but this at least gives us a little time to focus specifically on 
wrapping up the big ticket items before bug squashing ensues.

Suitable compromise for now?  Thoughts?

Thanks,

-b




On Tue, Aug 28, 2018 at 2:39 PM Kathy Lussier 
<kluss...@masslnc.org<mailto:kluss...@masslnc.org>> wrote:
Hi Bill,

In looking at the list of showstoppers, I see one has a pullrequest, so it 
seems reasonable it could be tested and merged soon. For the other bugs, does 
anyone have a sense of whether any are particularly complex? Or are they mostly 
straightforward bugs that just haven't been addressed yet due to lack of tuits? 
If it's the latter, could we consider delaying the full release (with xul 
removal) until the showstoppers are fixed?

I'm concerned about the breakage that is likely to occur the longer we continue 
to make the xul client available in our releases, but these bugs were 
identified as real issues in getting libraries to adopt the web client. At this 
time, there are just a handful of remaining showstoppers, and if we can commit 
to getting them resolved before the full release, I think it will make a 
smoother transition to 3.2 for our libraries.

Kathy

--
Kathy Lussier
Project Coordinator
Massachusetts Library Network Cooperative
(508) 343-0128
kluss...@masslnc.org<mailto:kluss...@masslnc.org>


On Mon, Aug 27, 2018 at 6:47 PM Bill Erickson 
<beric...@gmail.com<mailto:beric...@gmail.com>> wrote:
Hi Scott,

On Mon, Aug 27, 2018 at 5:24 PM 
scott.tho...@sparkpa.org<mailto:scott.tho...@sparkpa.org> 
<scott.tho...@sparkpa.org<mailto:scott.tho...@sparkpa.org>> wrote:
Hi Bill,
    I have two questions about this:


1.       You mentioned a vote. Who is the “we” that votes?

Good question.  This would be a core developer vote.   I started typing this as 
a developer list message, then added the general list just before sending...

>From my perspective, this vote is more about getting a public record of 
>developer buy-in (or otherwise) as is typically the case before proceeding 
>with a large architectural change.  It also acts as a "should we do this?" 
>safety valve.  However, I call the vote now because in my opinion as RM we are 
>ready to proceed and I suspect that's what we'll decide.  It's not done 'til 
>it's done, though.

It's also worth reminding everyone we are also providing extended support for 
Evergreen 3.1, so users can continue using the XUL client for a longer period 
of time.  Normally, a release is supported for 12 months of bug fixes, plus 3 
months of security fixes.  3.1 will be supported for a longer period of time -- 
duration TBD -- so sites will have more time before needing to upgrade to 3.2.  
This will buy us more time in the community to continue squashing bugs as well.

2.       If it is determined that not enough blockers are fixed, does this mean 
that a 3.2 version of XUL will be made available and XUL will not be removed 
until 3.3


Yes, if the core developers vote not to proceed with XUL removal, it would be 
delayed until the next release cycle (3.3).

Just to offer some perspective, from the dev side it's not just a question of 
how many web staff blockers remain, but how much work is required to resolve 
each, who can sign up to fix them, how many sites they likely affect, how much 
developer time will be siphoned away from fixing these issues trying to 
maintain XUL in 3.2 (!), the fact the XUL is already a little bit broken in 3.2 
based on the agreement it would it would be removed, etc, etc.

Thanks,

-b


Reply via email to