Jump to: navigation, search

CBI/Jan24 2012

< CBI

Meeting

  • January 24, 2012, 10am EST
  • Used new Eclipse Foundation conference facility (1-877-253-0615), Andrew's bridge

Attendees

 Andrew R., Igor F., David W., Kim M., Denis R., Markus K., Wayne B., Paul W.
 Regrets: Andrew O.

Today's agenda

  1. Quick FYI, new hire release engineer, Thanh Ha, joining Feb.27th
  2. Test results & bundle comparison see: http://build.eclipse.org/eclipse/e4/mavenCompare38M4/
  3. Known differences between CBI prototype build & PDE build
    • Add license feature support to Tycho, then license feature handling in the platform pom.xml's
    • Add source bundle handling to platform build
    • @qualifier@ replacement/handling in platform build (Bug 368497)
    • Signing

Not on agenda, still being tracked as work left to do

 a) Platform specific code swt & launchers
 b) Which execution environments will CBI build & test against
 c) maven.eclipse.org
 d) API Analysis
 e) Junit tests


Minutes

Step through Paul's results

Andrew & Paul stepping through Paul's test results to understand them a bit better.

 http://build.eclipse.org/eclipse/e4/mavenCompare38M4/

Igor: What is missing?

Paul: see this file

 http://build.eclipse.org/eclipse/e4/mavenCompare38M4/productNoSourceNoQualifier.diff.txt, named some specific examples
 org.eclipse.presentations (Igor: this is fixed already)
 org.eclipse.rcp_root
 org.eclipse.platform.doc.isv
 org.eclipse.equinox.slf4j.stub

Igor: Need to investigate to determine why the extra bundles were included for the CBI prototype build.

Andrew: I expect we should see these numbers match?: http://build.eclipse.org/eclipse/e4/mavenCompare38M4/productStats.txt

Paul: Confirmed, or get very close.

Andrew: What kinds of things affect this? Source bundles maybe?

Paul: Yes. That'll get them closer.

David: bug 369497 was wrong => Andrew: Was a typo, please see: 368497

Other differences noticed

Download sites

Andrew: the download sites aren't populated. What sets that?

Paul: Product controls the download sites, p2.inf

Igor: Suspect something broken in Tycho, should have had everything but doesn't.

ACTION: Igor, raise new bug to track this.

Other repositories

Paul: Andrew needs to set scope. e.g. equinox not being built thus repository numbers are much more different.

ACTION: Andrew

Operating systems

Igor: Operating systems like Sun, HP-UX, AIX can attribute. Out of scope for the prototype.

ACTION: Andrew track this as something the LTS Steering Committee will want to think about.

General

David: Does tycho use JDT compiler?

Igor: yes it does. It'll be updated to M5 as soon as it's out.

Totally off topic: Asterisk

This was the biggest call so far we've held on our new asterisk server.

For those that are curious, Asterisk bandwidth usage for our call with 8 participants.

Before the call:

 RX packets:890263 errors:0 dropped:0 overruns:0 frame:0
 TX packets:389372 errors:0 dropped:0 overruns:0 carrier:0
 RX bytes:117975247 (112.5 Mb)  TX bytes:118993294 (113.4 Mb)

After the call:

 RX packets:1515110 errors:0 dropped:0 overruns:0 frame:0
 TX packets:1015528 errors:0 dropped:0 overruns:0 carrier:0
 RX bytes:251232426 (239.5 Mb)  TX bytes:253972671 (242.2 Mb)
 

Thus 127 MB total or just under 16 MB per person. Based on the approximate length of the call (30 minutes), this works out to be about 550k/minute or 9k/second/person.

CPU & Memory usage hovered around 1% the entire time.

Thanks for those that gave me feedback that SkypeOut (my sound) was not great.