Notice: this Wiki will be going read only early in 2024 and edits will no longer be possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.
Attending: Oliver, Harm, Chris, AlexA, AlexN, Joanna, Paul
No objections to last week's notes.
I7 Status Update
- Paul reports that weekly schedule not updated last few weeks.
- The team is not on track to complete everything for iteration but will be ready to enter test pass with what is done.
- There are a couple of platform defects that impact test that we discussed
Joanna noted that Stanislav indicated that he cannot finish IPv6 items for I7.
- Stanislav reported IPv6 needs a few more weeks
- Jonathan had full doc that divided work between IBM part of code and Intel part of code.
- Mikhail noted that some critical AC defect from IBM got prioritized above IPv6 in previous iteration.
- It was known at the time that Intel did not have additional resource to bring to bear here.
- Because fixed size (0.5hc) maintenance resource being used for IPv6 in AC and core defects for AC, prioritizing any defects along with IPv6 put IPv6 at risk.
- Joanna wants to look forward from here to identify what is the remaining gap so we can figure out how to fill it (i.e., how much assist is needed to reduce risk on IPv6?)
- She is unsure who to talk to about what actually will be done on IPv6 and what is gap
- Mikhail will look into it.
- Goal: by end of week understand what will be completed and what is remaining gap to allow action plan to be developed.
The other enhancement area where IBM is concerned about finishing for 4.5 is the Java 6 StackMap attribute patch.
- Chris gave an update on where this is. The version approved in IPzilla (Harmony milestone 5) had some boundary condition issues that we found during integration. Harmony team with TPTP guys identified fixes and have implemented a solution.
- Request in to Harmony to accelerate milestone release w/ fix (a milestone 5.5)
- Request into IPzilla to get it approved. They agreed to try to rocket it thru.
- Mikhail mentioned that the Harmony team states they are nearly done with making this happen.
- Compiler compatibility between compiler used by Harmony and TPTP is remaining question
- Will see if can switch from "unmodified binary" to "unmodified source" so we can use TPTP compiler
- Harmony planning milestone 5.5
- Code hoping to freeze next week w/ release about a week later
- Question from Mikhail whether we can take code upon freeze and submit for review or if we need to wait till after formal release.
- Chris noted that for 5.0 approval in Feb, we submitted to IPzilla around the time of the freeze in order to expedite approval. This should be okay if we are convinced the code is really done.
Joanna noted that 4.5 I7 has some accessibility defects for AlexN to look at.
AlexN mentioned that there is risk because the team could not get all the P1s done.
- They have a bigger backlog now.
- Some stuff will have to slip out of 4.5.
- Non-critical P1s are top of list for P8. He needs to prioritize stuff that will slip out of 4.5
- Oliver asks if there are any major questions to raise to PMC about this at this time.
- Not right now
AlexA gave an update on I7 defect status.
- No issues in this iteration except the bug we talked about already in Platform section (Harmony verifier)
- Team is finalizing most of the bugs for this iteration
[Note: Chris was having phone issues and dropped off for a bit of the discussion]
Post 4.5 Resourcing
Intel has gotten some guidance from management regarding post 4.5 participation. Intel will be committed to maintenance requirements after Ganymede release. It is likely that some new contributors from Intel will become involved.
- Oliver expressed his concerns about how he had been counting on at least a flat investment level. This announcement differs from the feel that he got from discussions around EclipseCon.
- Harm has concerns around features that have been put in recently to ensure support is effective
- Examples: Threading updates, binary transfer format, etc...
- These are new and we expect that will have some hardening to do
- Joanna expressed concerns around whether AC is included in support plans
- Chris indicated that his gut instinct is that there is some level of support obligation around both profiling and AC
- Harm asks about transition overlap between existing resources and any new resources that may come on board
- It was too early for Chris to comment on this
- Harm noted (paraphrasing) in the past overlap was important to ensure successful transition
Oliver noted that a reduction to maintenance mode would effectively eliminate the POG effort
- Chris noted that while additional hardening of existing POG use case will happen that new usability enhancements for other use cases would be a bigger challenge.
- Harm noted that we should reconsider being part of the WTP JEE EPP bundle
- It was in essence a publicity stunt for POG to attract new users and therefore use cases.
- This resourcing change suggests that we need to be more conservative to avoid misleading potential users.
- In particular Harm asks if being part of EPP makes us take a bigger risk than we need to.
- Joanna agreed
- Oliver agreed
- Chris could go either way. He thinks that the existing use case that has already been hardened significantly will be of value to EPP users but realizes that it could set up some expectations that would be difficult to follow thru on.
The PMC agreed to look into NOT being part of EPP
- Harm would like to know sooner rather than later
- AI for Joanna to touch base w/ IBM sponsors to make sure its okay to pull out of EPP from their perspective.
Oliver would like to chat a bit more with Intel management about the changes
- AI Chris to try to set up discussion with next line.
Next steps for PMC
- We will need to embrace IBM more for forward looking plans since they are the last remaining investor of size...
- Oliver asks what does health of project look like if only IBM is contributing enhancement resources
- AI to Joanna/Harm to come back with w/ detail about how IBM will react