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.
Difference between revisions of "Equinox/p2/Meetings/20090914"
(→Attendees) |
(→Minutes) |
||
Line 16: | Line 16: | ||
* Profile comparison in installation history | * Profile comparison in installation history | ||
* Introductory p2 webinar tomorrow (12pm EST) | * Introductory p2 webinar tomorrow (12pm EST) | ||
+ | * Moving p2 to 1.5 | ||
+ | ** The move to 1.5 will only be done if we can ensure down compilation | ||
+ | ** We also discussed the potential overhead of Collections for objects like RequiredCapabilities | ||
+ | ** At this point down compilation requires to not use enums, and not use any language construct that would relies on Iterable. | ||
+ | ** Things to investigate: | ||
+ | *** Are there any additional constructs to avoid? | ||
+ | *** Does javac allows for down compilation? | ||
+ | *** Can code compiled with Eclipse be compiled against with Javac? | ||
+ | *** Does PDE Build need to be changed? | ||
+ | *** Will API tooling help us catch wrong library usage? | ||
+ | |||
+ | * DI exploration | ||
+ | ** We have had a long discussion about our a DI approach to p2. | ||
+ | ** In order to get a better understanding of the problem, we will make the changes in HEAD with the main goal of changing the director application. | ||
+ | ** Henrik has a context mechanism that we will see if it can be applied. |
Revision as of 22:08, 14 September 2009
Attendees
- Andrew Niefer
- DJ Houghton
- Ian Bull
- Jeff McAffer
- John Arthorne
- Pascal Rapicault
- Simon Kaegi
- Tom Watson
Minutes
- Test pass tomorrow
- New SAT4J
- Profile comparison in installation history
- Introductory p2 webinar tomorrow (12pm EST)
- Moving p2 to 1.5
- The move to 1.5 will only be done if we can ensure down compilation
- We also discussed the potential overhead of Collections for objects like RequiredCapabilities
- At this point down compilation requires to not use enums, and not use any language construct that would relies on Iterable.
- Things to investigate:
- Are there any additional constructs to avoid?
- Does javac allows for down compilation?
- Can code compiled with Eclipse be compiled against with Javac?
- Does PDE Build need to be changed?
- Will API tooling help us catch wrong library usage?
- DI exploration
- We have had a long discussion about our a DI approach to p2.
- In order to get a better understanding of the problem, we will make the changes in HEAD with the main goal of changing the director application.
- Henrik has a context mechanism that we will see if it can be applied.