Minutes of Edinburgh Release Meeting, 08 Jan 2007

TOC and Agenda

Logistics, Participants, IRC log

• Date: Monday, 08 Jan 2007 21:00 GMT
• IRC Channel: #twiki_edinburgh
• Participants: KennethLavrsen (Lavr), PeterThoeny (PeterThoeny), HaraldJoerg (HaraldJoerg), SvenDowideit (SvenDowideit), CrawfordCurrie (CDot), SteffenPoulsen (SteffenPoulsen), StephaneLenclud (Stephane Lenclud)
• Not present but usually are: ThomasWeigert (tweigert), KwangErnLiew (ktwilight), AndyGraybeal (AndyGraybeal), Benjamin Jones (Dzaster), TravisBarker (TravisBarker), MartinCleaver (MartinCleaver), OliverKrueger (OliverKrueger), WillNorris (wbniv), MichaelDaum (MichaelDaum), SamHasler (Flenser), RafaelAlvarez (Soronthar), AntonioTerceiro (terceiro), ArthurClemens (ArthurClemens), LynnwoodBrown (Lynnwood)
• Facilitated by PeterThoeny (PeterThoeny)
• Notes taken by KennethLavrsen (Lavr)
• Meeting Minutes: EdinburghMinutes2007x01x08.txt

1. Review Previous Action Items

• Skin experts. Give Thomas final feedback on TemplatePathBuginTWiki4x00 so he can finish the docs
  • It seems the release meeting participants do not really understand what feedback is needed. Any further actions goes beyond 4.1. Thomas to continue discussing on Codev topic.
    • Just as a reminder what I was hoping for: There have been signficant changes to the template lookup algorithm in Dakar. Some of these changes seem unmotivated and make the algorithm difficult to explain. The key question is: Should the algorithm be documented as is, or should it be analyzed whether the strange additions were really warranted? (Look for example, at steps 1 and 7 in the Dakar version.) There is a recommended algorithm there. I was hoping to get feedback on that. All this is at the very top of the page.
    • Further, the current documentation does not exactly match the algorithm, and also states that there are certain items deprecated. I personally am fine with the deprecation, but did that really happen? Again, I'd like feedback as to whether we should leave the slightly incorrect current description in order to allow to leverage the ambiquity to later fix the algorithm?
    • I make some minor stylistic changes to the documentation. If there is no further feedback, I will leave the algorithm documented as is, that is omitting some steps and leaving ambiguity as to how far step (5) in the documented algorithm extend (this corresponds to step 7 in the Dakar version).
    • My suggestion is the algorithm marked as recommended in TemplatePathBuginTWiki4x00, possibly removing the steps marked as deprecated in the current documentation. -- TW
• Peter: Skin/Javascripts changed so initial letter for renaming and creation changes first letter to uppercase. Classic skin changes are pending
• Bug assignements:
  • Bugs:Item3261 Literal search on all webs does not work. Peter will try and fix
  • Bugs:Item3370 Release note for 4.1
    • Kenneth finishes the release note
    • Peter adds a contribution list on TWikiHistory and updates AUTHORS file.
  • Bugs:Item3371 Walkthrough of all default plugins to ensure latest version of default plugins are on TWiki.org. Kenneth will start this.
  • Bugs:Item3378 default templates cause invalid HTML
  • Bugs:Item3399 Template system recursion prevention too agressive, prevents skin specialisation, or mixins
    • This item is not a requirement, but an enhancement request. I think one week before a release this is way too late. There has been no discussions, so the 4 checkins on this item should be reverted. -- TW
    • TW is incorrect. TWiki-4.0.* contains this functionality and I know of a number of sites that use it. It is THE bug that I was refering to months ago when TW was working on the templates, and i was shocked to see it was not fixed. -- SD
    • Just to set the record straight:
      • TWiki-4.0 did not have a check against recursion inclusion.
      • Recursion prevention was introduced by Sven in Bugs:Item2530. The bugs items literally states that this is meant to exclude calling %TMPL:INCLUDE{"view"}% from view.svenskin.tmpl.
      • However, that version ignored inclusion from user topics, which was added in my code.
      • The recent change is meant to specifically allow the situation that the earlier bug item tried to prevent.
    • Meeting agreed this is fixed and only action is for SD to explain to TW (the test case that has now been added is sufficient explanation -- TW )
      • Note that in the IRC there is some claim that SKIN=myskin,pattern does not work, but that is not true in general. I am using this all the time, but not in the recursive version. I looked at the test case that Sven finally added. Pretty slick usage, but I am sure that was not intended originally... Anyway, looks like we have a decision...
      • Sorry Thomas, I disagree. This was the intended original functionality, one that I originally used in Cairo, and re-implemented in January2006 for a contract I had, and then finally remembered to commit to the core when I found out it also happened to stop recursion. The entire body of template and skin work i'm doing are slowly bringing to TWiki what I envisioned (and used in private branches) a long time ago :/ -- SD
        • While I fail to see how the presence of a feature in a private branch indicates that this was the original TWiki intention, I think we should stop arguing over this as (i) we have a decision, and (ii) this is a good enhancement in any case. -- TW
  • Bugs:Item3401 base tag refers to topic view, irrespective of what the current base url actually is
    • Needs someone to action
  • DropClassicSkin is unloved
    • This is an ongoing discussion which I also participate in but the scope has never been 4.1. We are not going to rip out a skin 1 week before release. KennethLavrsen.
    • good - its just you mentioned this meeting in that topic.. -- SD
    • Yes. I suggested that we add a deprecation note in release note. Sorry forgot. -- KJL
    • Decision:
      • 1. no deprecation
      • 2. keep default templates and classicskin out of sync for 4.1
      • 3. derive classicskin from default templates in a patch release
  • See also "top candidates of normal bugs items" at EdinburghReleaseMeeting2007x01x02

2. Review Bug Fix Status of TWiki 4.1

• Desired outcome: Decide which non-blocking bugs should be fixed

• Bugs:Item3393 sounds like it might be a potential security hole; needs to be checked by Peter

3. TWiki 4.1 Release Coordination

• Desired outcome: Coordinate tasks of ProductionReleaseChecklist

• From previous meeting:
  • Release date has been defined as 14th of January 2007.
  • KennethLavrsen builds the release.

• Decisions:
  • Based on votes in CodeName topic, the next codename is Freetownrelease
  • Kenneth will keep the TWikiReleaseManagerRole until at least 4.1.1

Action Items

• Kenneth to remove upgrade script and change upgrade doc so only manual process is left in. There may be libs also. The remove is in MANIFEST only.
• Item3401 needs someone to action it.
• Kenneth adds to release note that in 4.1.1 the FuncContrib stuff gets merged to func. Bugs.Item2354
• Kenneth makes RC3 Tuesday evening.

IRC Log

• EdinburghMinutes2007x01x08.txt: Edinburgh Meeting Minutes 2007-01-08

Back to: EdinburghRelease

Comments / Feedback

Fixed typo in release date (was 2006, should be 2007).

-- JoachimBlum - 09 Jan 2007