Notice: This Wiki is now read only and edits are no longer possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.
DSDP/TM/Committer Phone Meeting 11-Sep-2007
| Meeting Title: | TM Committer Phone Meeting |
| Date & Time: | Tuesday Sep 11, 2007 at 1500 UTC / 1100 Eastern |
| Dial-in: | Martin to call everybody by Skype |
Backup dial-in: International +44 (0)1452 567588 / Freephone +1 (866) 6161738 / UK 08712460713 / Passcode: 0587322148 #
MartinO to start conference call - please dial in using the numbers above.
Please be available for Skype Chat in parallel to the call. MartinO will start Skype chat just prior to call.
Skype dial-in: martin.oberhuber, ddykstal (or david_dykstal), david-k-mcknight, kushal.munir, kevin.j.doyle, xuan.chen886, rupen.mardirossian, javier.montalvoorus, tedatteddotnet, michael_scharf, and uwe.stieber.
Attendees
- IBM - Xuan Chen, Dave Dykstal, Dave McKnight, Rupen Mardirossian, Kevin Doyle
- Wind River - Martin Oberhuber, Uwe Stieber, Michael Scharf
- Symbian - Javier Montalvo Orus
This is an Open call, so anyone else can join (though we expect the talk to be interesting for committers only).
Agenda
- Last meeting: DSDP/TM/Committer Phone Meeting 5-Sep-2007
Current Work
- Skype Call Quality
- Intensive TM 2.0.1 Testing on Thursday
- Reminder: Aug 30th was RC1 as per TM 2.0 Ramp down Plan for Europa#Ramp_down_for_Europa_SR1_.2828-Sep-2007.29, no risky fixes for 2.0.1 after the 30th!
- Bugzilla TM "3.0" target milestone has been created; creating also "2.0.2" milestone; note that 2.0.2 bugs MUST be checked in to the branch, which is some effort, so this should be reserved for really important things only
- Branch not yet forked off, but policies do apply (need review)
- OK to check in to HEAD, Martin can merge to the branch
- Need 1-committer review, and Verification of fixes in an I-build after that time (RC1)
- Reminder: Start preparing / planning for the DSDP/TM/Face-to-face Meeting Toronto 17-Sep-2007
- Room with outside internet access should be available in customer briefing center
- Xuan: How to prepare?
- 1. Make up your minds WHAT you want to achieve for 3.0; we should come up with plan items, priorities, and target milestones
- 2. Get initial ideas how to achieve this, such that we can discuss architecture and approaches.
- What's a Coding Camp? - Martin: We'll do a planning meeting, but also with hands on the code for prototyping and experimenting
- DaveD: Persistence, User-Defined Actions - Explain the plans, discuss them
- DaveM: Need to think
- MartinO: UI/Non-UI; New-connection-wizard: more user friendly while keeping it pluggable
- Kevin will be dedicating some time (up to 20 hrs a week)
- Rupen will be continuing (tentative)
- DaveD: -
- bug 194802 - NPE when restoring; might be gone in the meantime, analyze / reproduce? - closed WORKSFORME
- What to do in the future? - Martin: put a note on bug 202866, in the future we should move all SystemRegistry to non-UI and ensure that whatever client gets a SystemRegistry already gets an initialized one (i.e. synchronous initialization, like Eclipse core.resources). But we can discuss this in the future.
- bug 198395 - Connect dstore with expired password - need to look at perl APIs for pwd access - not for 2.0.1
- bug 197036 - Avoid plugin activation, dont create all filterpools on startup - not for 2.0.1
- bug 176577 - Moving hosts up+down - looking at for 2.0.1 today
- bug 194802 - NPE when restoring; might be gone in the meantime, analyze / reproduce? - closed WORKSFORME
- DaveM: - looking at these this week
- bug 196662 - Refresh queries on dispatch thread - initial problem fixed but not ideally; some additional changes will be needed but they need API related to bug 187739 and an exists() method in the SystemViewElementAdapter -- defer after 2.0.1
- bug 200557 - Review dnd from Project Explorer (bug 192704 ok
- bug 199565, bug 199566, bug 199568 - Deadlock issues due to synchronized methods calling out; will be able to fix
- Kevin: - bug 168219 recursive delete w/read-only moved to "Future"
- Xuan: - bug 202686 - going to disable or fix supertransfer? - Found solution for two issues, a third one (refresh) is still open due to race condition (works during debug but fails in nondebug); if unable to find the issue, will disable supertransfer in the Preferences.
- Martin: Can use java.io.File#deleteOnExit() in order to make the temporary zip deleted on disconnect dstore;
- Xuan: Seems to be more a view update issue than actually on the Server
- Javier: - will look at bug 199773 FTP upload as binary
- Found a problem with UI/Non-UI splitting, because Ascii/Binary properties are living in UI -- FTPService gets wrong isBinary tag (DaveM: DStore - RemoteFile#isBinary() seems to be OK) - Javier will further work on it
- bug 202758 FTP hangs completely due to missing server answer for completePendingCommand() - not reproducable by Martin, but reproducable by Michael; Michael writing a Unittest
- Martin: - bug 202630 what is the meaning of the "default private system profile"? - WR Workbench automatically fills this from a Registry. Should RSE ensure that there is always exactly one default private profile? Into what profile should connections be auto-generated... should there be a "WR Registry" profile? But then, what if users create connections themselves, would they create them into the "WR Registry" profile?
- DaveM: Original design was that there is always exactly one "default private" profile, which is not intended to be shared;
- DaveD: The notion of "default private" is legacy -- all profiles should be inherently shareable; default private is a persisted attribute, but probably should not be (Martin is OK with having a persisted attribute "private" - but the attribute "default" can not really be persisted because we don't know who claims to default)
- Martin: ensureDefaultPrivate() is not consistent with getDefaultPrivate() -- if it already finds an active default private one, it should not try creating a new one;
- Will need to think about how to address in our product
- bug 202416 NPE when loading profile: made it more robust, but how could it come that RSE profile data is wiped with zeros? What happens when an IOException occurs while saving a profile, can it become corrupted or is there a safeguard?
- bug 202866 Exceptions when browsing during early init - DaveM please review
- Uwe: -
- Rupen: -
- Michael: - Terminal Performance branch: 300k lines/sec on Windows, 8000 lines/sec on Linux; everybody is OK with putting this into 2.0.1
- Questions
- Bugzilla:
Vacation, Away
- DaveD vacation Spt 27 - Oct 9
- MartinO mostly away Sep24 - Sep 28 (only doing releng stuff that week)
Action Items
- Last Meeting Action Items
- DaveD: 2.0.1 fixes, then Doc bugs (Tutorial)
- DaveM: 2.0.1 fixes
- Xuan: 2.0.1 fixes (Supertransfer)
- Kevin: testing
- Martin: Look at PropertyDescriptor issues; EFS fixes; unit tests; Migration Docs, Releng Fixes, Newsgroup
- Javier: 2.0.1 fixes, unit tests
- Michael: Terminal performance improvements