ICU Meeting Minutes 2009 (Archived)

Current and past meeting minutes

2009/12/30 no meeting

2009/12/23 cancelled, only 2 participants

2009/12/16

(Phone conferencing system was down for 8AM PT CLDR mtg, seems fixed for this meeting)

  1. Incoming bugs
  2. Very quick status on normalization (Markus)
    near implementation complete on C++ side, working on C API
    plan to merge end of week then work on J.
    follow-up work after API freeze: replace remaining old implementation code, truly load & cache non-standard data files, interesting test data, test API & increase coverage
  3. API freeze date (Markus)
  1. what does it mean to have API freeze by Jan 13?
  2. want to check in API through end of 2nd wk Jan
  3. move milestone out by a wk - last checkin Jan 15
  4. milestone M4 jan 20th with API freeze.
  5. milestone M5 feb 24th has code freeze, also picks up feb 10 CLDR final candidate
  6. this allows a short time for client test integration & testing before mar 10 release
  1. Relationship of Olson data to geonames database (Peter)
  1. san luis vs buenos aires - error?
  2. The main 'geoname' table has the following fields :
    ---------------------------------------------------
    geonameid : integer id of record in geonames database
    name : name of geographical point (utf8) varchar(200)
    asciiname : name of geographical point in plain ascii characters, varchar(200)
    alternatenames : alternatenames, comma separated varchar(5000)
    latitude : latitude in decimal degrees (wgs84)
    longitude : longitude in decimal degrees (wgs84)
    feature class : see http://www.geonames.org/export/codes.html, char(1)
    feature code : see http://www.geonames.org/export/codes.html, varchar(10)
    country code : ISO-3166 2-letter country code, 2 characters
    cc2 : alternate country codes, comma separated, ISO-3166 2-letter country code, 60 characters
    admin1 code : fipscode (subject to change to iso code), see exceptions below, see file admin1Codes.txt for display names of this code; varchar(20)
    admin2 code : code for the second administrative division, a county in the US, see file admin2Codes.txt; varchar(80)
    admin3 code : code for third level administrative division, varchar(20)
    admin4 code : code for fourth level administrative division, varchar(20)
    population : bigint (4 byte int)
    elevation : in meters, integer
    gtopo30 : average elevation of 30'x30' (ca 900mx900m) area in meters, integer
    timezone : the timezone id (see file timeZone.txt)
    modification date : date of last modification in yyyy-MM-dd format
  3. Peter to investigate more, make a proposal for more integration and/or testing
  1. Changes in for tickets 7263727972807284; might be worth checking for correct behavior on different compilers (Peter)
  1. check trunk
  2. srl to get config results to markus after jan 15, 2010
  1. 'formatToCharacterIterator' API for ICU4C (doug).
  1. add overrides to take fieldposition -> fieldposition iterator
  1. meeting schedule
  1. Dec 23, 2009 - srl will be in, small meeting
  2. Dec 30, 2009 - NO Meeting
  3. Jan 6, 2010 - meeting.

2009/12/9

  1. C++ substrings - defer, need more internal discussion (Markus)
  2. Normalizer2 API for custom normalization (Markus)
  1. API proposed yesterday
  2. No replacement proposed for iterative normalization and normalization string comparison; use existing API for working with standard Unicode normalization
  3. Java would use CharSequence & StringBuilder
  4. C++: may want to use new SubString class for input
  5. getInstance() for custom data, provide data how? InputStream? Look at StringPrep (.spp StringPrep profile) and DBBI (.ctd dictionary)
  1. CLDR 1.7.2 (Yoshito)
  1. LDML 1.7.2 final draft - http://unicode.org/draft/reports/tr35/tr35.html
  2. CLDR 1.7.2 release / end of Thursday?
  1. Currency data API semantics (Doug)
  1. return null and document for getAvailableCurrencyCodes
  2. document that order is significant and reference CLDR section on ordering
  3. when getKeywordValuesForLocale and commonlyUsed = false, order is not significant, if true, order is preference order
  4. if specific region, no data, return empty array
  5. if no specific region, use likely subtags, and then do as above
  6. don't flesh out 'und' because it will get US for the region
  7. if keyword is not "currency" return empty (we don't recognize the keyword)
  1. Incoming bugs

2009/12/2

  1. ICU4J 4.4 modularization (Yoshito, Doug?)
  1. What in official release?  Still big all-in-one icu4j.jar (+ charset, localespi) -> we do not change release jar files
  2. What to do for existing module jar support? -> No changes in 4.4, but tell our users that this will be removed after 4.4
  3. Any way to describe module dependencies? (probably future topic)
  1. OSGi bundle - of course, it requires OSGi framework, but Yoshito's understanding is that it just require extra headers in manifest - so not harmful
  2. Jigsaw/JSR-294 - static module dependencies can be described in the source code, but require Java 7
  1. C++: There is some interest (Google, Apple, Harfbuzz) in splitting DLLs into smaller chunks. Maybe after 4.4.
  1. DateFormatPatternGenerator options (Peter)
  1. Most recent e-mail, "Re: [icu-design, cldr-dev] DateTimePatternGenerator option to force digit length match", Nov 24, 10:35 AM PT
  1. UnicodeString DateTimePatternGenerator::getBestPattern(const UnicodeString &, UErrorCode &, DTPGenOptions options = DTPGEN_NOOPTIONS); Don't do this. Use overloads as on the next line. -Markus Scherer 12/2/09 11:17 AM 
  2. UnicodeString DateTimePatternGenerator::getBestPattern(const UnicodeString &, DTPGenOptions , UErrorCode &);
  3. DateTimePatternGenerator::setOption(...); -- if the generator is mutable. <- It is not, so we go with #2.
  1. Duration formatting (Peter)
  1. Discard current data approach? Yes, instead need new structure in LDML for this, Peter to file CLDR bug (e.g. to get localized duration such as hh:mm:ss)
  1. Normalization (Markus)
  1. design doc: http://site.icu-project.org/design/normalization/custom
  2. new API proposed, can load custom data
  3. out of the box: nfc, nfkc, nfkc_cf (Unicode standard: with case folding and removing default ignorable code points)
  4. may add data for UTS #46 (Unicode IDNA Compatibility Processing)
  5. old API will be routed to new one, use cached data (lazy load)
  1. BCP 47 -u- spec (Yoshito, Mark)
  1. CLDR 1.7.2 - only BCP 47 related spec/data changes
  2. Will be ready for review - next Tuesday?
  3. -u- for canonical representation for Unicode language/locale identifier - @syntax for backward compat in LDML spec
  4. en_US_POSIX -> en-US-u-va-posix
  1. Incoming bugs
  1. nonworking URL: http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&priority=%2521zero&owner=&owner=somebody&order=id

                

   


Ticket

Summary

Priority

Status

Owner

Type

Milestone

7261

fix uintrnal.h generation by genheaders.pl

major

new

somebody

defect

4.3.4

7263

C headers fail when included inside extern "C"

assess

new

somebody

defect

UNSCH

7267

ICU4J FAQ looks like it hasn't been updated in 10 years

assess

new

somebody

task

UNSCH

7269

uloc_getCountry and uloc_getVariant not working properly.

assess

new

somebody

defect

UNSCH

7270

Line Break: implement revised LB8 rule

assess

new

somebody

defect

UNSCH

7279

Need a way to set the default value of U_HIDE_INTERNAL_API

assess

new

somebody

defect

UNSCH

2009/11/25 - no meeting

2009/11/18

  1. Introducing Michael Grady (Apple - Peter)
  2. Milestone 3 status (yoshito/michael)
  1. Milestone 3: review API change reports (Yoshito/Markus)
  1. ICU4C LCID mapping (yoshito/michael)
  1. http://www.icu-project.org/trac/ticket/7244
  1. root locale and locale fallback etc. (yoshito/mark/markus/doug/umesh)
  1. http://site.icu-project.org/design/locale/root
  1. remove old way of building ICU4J data jars from ICU4C -- remove gendtjar.pl? (Markus)
  1. Display name API (doug)
  1. Doug to schedule a call discussing this topic in sub group.
  2. (Meeting Document)
  1. Incoming bugs

2009/11/11

  1. Milestone 3 status
  1. Friday - final checkin, communicate with Yoshito/Michael if you have anything after that
  2. Unicode 5.2 integration
  1. C completed
  2. J need some test case changes
  3. J Normalizer bug?
  4. Hopefully everything wrapped up by Friday - might be delayed..
  1. CLDR using /dev/ - Mark to generate list
  2. UTrie2 to Java (Andy) - almost working
  1. casing/inlist (jason)
  1. Rietveld for ICU code reviews (Markus/Yuqian)
  1. ULocale#getFallback - String/ULocale
  1. http://www.icu-project.org/apiref/icu4j/com/ibm/icu/util/ULocale.html#getFallback%28java.lang.String%29
  2. http://www.icu-project.org/apiref/icu4j/com/ibm/icu/util/ULocale.html#canonicalize%28java.lang.String%29
  3. "en__POSIX" -> "en_" -> "en" -> "" (current behavior)
  1. "en__POSIX" -> "en" -> "" (new behavior)
  1. "en" -> "" (current behavior) #6673
  1. ULocale.ROOT = new ULocale("root")
  2. ULocale.ROOT != new ULocale("")??
  3. "und" ???
  4. What should the canonicalize() method do?
  1. Consensus to canonicalize root, und, empty to same thing
  2. Consensus to canonicalize all three to the empty string; bundle lookup will still load root.res
  1. What should constructor do?
  1. Should/does adjust for casing?
  2. Does not normally canonicalize, but should it canonicalize these three special locale IDs?
  1. What should be the value of ULocale.ROOT? "root" vs. ""
  2. C++ vs Java? - Should behave the same
  3. Resource-Bundle lookup? - need to continue to load root.res
  1. John's change?: en_Arab_CA => en_Arab => root
  1. 2009r update (yoshito) - needed metazone update, CLDR 1.7.1.2
  2. Incoming bugs

2009/11/4

  1. Incoming bugs
  2. casing/inlist (jason)
  1. Still waiting for the conclusion in CLDR
  1. locale data split (doug)
  1. icu4_3_2_milestone tag in CLDR?
  2. re-sync with CLDR
  3. ICU4J work
  4. Targetting M3
  1. plugin loading (srl)
  2. CharsetEncoder(UTF-16BE).maxBytesPerChar()
  1. m3 status (steven/yoshito)
  1. Unicode 5.2 status?
  1. Markus: ICU4C branch is working
  2. Started ICU4J branch for working on this
  3. Targetting M3 / may need someone's help
  1. Smart pointer (markus)
  1. Already working in Markus's branch

2009/10/28

  1. Incoming bugs
  1. Platform List for 4.4
  1. IBM:
    C regularly tested:
    Adding Windows 7 32 bit
    Removing AIX 5.2
    Removing Win2000
    Remove: Fedora FC10
    Remove: FreeBSD
    Add: Haiku
    Add: IBM i
    Add: IBM z

    J: adding OpenJDK 6 and 7 
  2. we should expand coverage for different GCC versions.
  1. 2009p breakage (Yoshito, Andy)
  1. http://www.unicode.org/cldr/trac/ticket/2444
  2. Yoshito, planned fix EOD tomorrow 10/29/2009

2009/10/21

  1. Incoming bugs

2009/10/14

  1. CLDR data not accessible via ICU API, spreadsheet in progress; feel free to add missing info: ICU use of CLDR Data. (Peter)
  2. C++ support for templates? (Markus)
  1. M3 schedule (Markus)
  1. Rolling in Eric's lenient parsing changes (Peter), e-mails of Oct 9
  2. Incoming bugs
  3. Party status- if you're within 100.000mi of SJ - respond
  4. Unicode Haiku
  1. ユニコード
    雲丹と鱈子で
    うま ご飯

2009/10/07

  1. 4.4M2
  1. SelectFormat (Norbert)
  1. Sync ICU & CLDR milestones (Markus/Mark)
  1. Not everyone uses stock ICU data (e.g., consensus to remove Unihan tailorings, but some users may need them; or choosing different alt="..." variants) (Mark/Markus)
  1. Need to be able to generate ICU data from CLDR for a milestone, with CLDR tools able to build against ICU and with CLDR data matching what went into ICU
  2. John tagged CLDR for ICU 4.4M2 (4.3.2): http://unicode.org/cldr/trac/browser/tags/icu432milestone
  3. Listed on BRS? - Markus updated the BRS list
  1. C++ template code: LocalPointer & MaybeStackArray (Markus)
  1. Party
  1. Restaurant, @ Santana Row
  2. http://www.straitsrestaurants.com/index.php?section=13
  3. Final count
  1. Incoming bugs

2009/09/30

  1. 4.4M2
  1. Final check-in by Oct 2
  2. Any changes after Oct 2 should communicate with Steven/Yoshito
  1. SelectFormat (Norbert)
  1. [icu-design] SelectFormat enhancement for MessageFormat (archive)
  2. default case: "default" or "other" (as in pluralformat)
  3. comment on the icu-design list. Norbert to write a reminder with a deadline.
  4. (note:)
  1. API slush Nov2009  (at least stub APIs checked in), API freeze Jan2010
  1. ICU 4.6 (CLDR 1.9) plan (Yoshito)
  1. CLDR meeting (9/30) minute :
  1. Tentatively ICU 4.4.1 in fall 2010
  1. Focused on U6.0
  2. CLDR 1.8.x release ( no new data )
  1. ICU 4.6 / CLDR 1.9 Mar/Apr 2011
  1. Unicode 6 isn't just a minor update- new APIs, etc.
  1. Probably more appropriate to call it 4.6 than 4.4.1
  1. U6.0 depends on ISO ballots/discussions. Not guaranteed. Will know by May 2010.
  2. Treat 4.6 as if it was a 4.4.x release (but call it 4.6 because it has new API (enum) and behavior changes).
  1. copy maint-4-4 to maint-4-6  in ~Aug timeframe (to start Unicode 6 beta)
  1. -> effectively,  4.4  + CLDR 1.8.x + Unicode6 + (cherry-pick from 4.8/trunk)
  2. Ok to cherry-pick post-4.4 features & fixes if we can manage the process
  1. trunk will be 'open' for 4.8 right after 4.4 release
  1. Move some tools out of the "icu" tree of repository and into the "tools" tree? (Markus)
  1. Party
  1. http://icu-project.org/party10
  2. Wed 10/14  6:30p @ google (main quad/HQ) - catered food
  3. Srl will contact Mark offline
  1. Incoming bugs

2009/09/23

  1. CLDR data update into M2? John
  1. infrastructure changes for CLDR, structural changes, ..
  2. want to tag M of CLDR with each M of ICU, at least tagging tools and data.
  3. will be on a branch
  1. Data Duplication (zone.tab and CLDR) John, Yoshito
  1. America/Los_Angeles = territory "US" ( CLDR zoneFormatting )
  1. cldr
  1.             <zoneFormatting multizone="001 AQ AR AU BR CA CD CL CN EC ES FM GL ID KI KZ MH MN MX MY NZ PF PT RU UA UM US UZ" tzidVersion="2009f">
                            <zoneItem type="Africa/Abidjan" territory="CI"/>
                            <zoneItem type="Africa/Accra" territory="GH"/>
                            <zoneItem type="Africa/Addis_Ababa" territory="ET"/>
                            <zoneItem type="Africa/Algiers" territory="DZ"/>
                            <zoneItem type="Africa/Asmera" territory="ER" aliases="Africa/Asmara"/>
                            <zoneItem type="Africa/Bamako" territory="ML" aliases="Africa/Timbuktu"/>
                            <zoneItem type="Africa/Bangui" territory="CF"/>
  1. zone.tab
  1. #code        coordinates        TZ                        comments
    AD        +4230+00131        Europe/Andorra
    AE        +2518+05518        Asia/Dubai
    AF        +3431+06912        Asia/Kabul
    AG        +1703-06148        America/Antigua
    AI        +1812-06304        America/Anguilla
    AL        +4120+01950        Europe/Tirane
    AM        +4011+04430        Asia/Yerevan
    AN        +1211-06900        America/Curacao
  1. US = America/New_York, ... America/Los_Angeles ( Olson zone.tab, in zoneInfo.res )
  1. Size of zonestring (Andy)
  1. parsing- can messageformat parse "{0} Time" ? - more interesting case - "Time of {0}"!
  1. Party!!
  1. Committers list - manual spreadsheet:  —>>  [   ____________   ]
  2. Time
  1. tie between Wed and Thu?
  2. Wed- opening reception (10/14)
  3. Thu- Conference reception (10/15)
  1. Place
  1. at Google - close to conf.
  1. Stuff
  1. IBM (snacks)
  2. Refreshments may come with the room?
  3. Party favors?
  1. ¡¡Party!!
  1. https://spreadsheets.google.com/viewform?hl=en&formkey=dEVEeDJUWEJTMEY4XzFIVHdpclZ5Z3c6MA
  1. << http://icu-project.org/party10 (redirect link)
  1. Wednesday oct14 - unless we get more votes.
  2. https://spreadsheets.google.com/ccc?key=0AqRLrRqNEKv-dEVEeDJUWEJTMEY4XzFIVHdpclZ5Z3c&hl=en
  1. Ohloh.net
  1. https://www.ohloh.net/p/icu/contributors
  2. https://www.ohloh.net/p/8372/contributors
  3. https://www.ohloh.net/p/cldr/contributors
  4. https://www.ohloh.net/p/cldr
  1. Incoming bugs

2009/09/16

  1. Schedule
  1. M2
  1. won't have Unicode 5.2
  2. minimal BRS
  3. Checkin by Oct 2
  4. release oct 7, 2009
  1. M3
  1. Nov 4, 2009
  1. Please send your configure'd platform.h and config.log to Markus
  1. "Re: [icu-design] ICU4C: un-autoconf platform.h?"
  2. Work incrementally, easy switches first
  3. Consider only one makefile configure'd and then included by static (not configure'd) per-folder makefiles
  1. Parsing Negative Numbers In sv_SE (Mark - see email)
  2. Incoming bugs

2009/09/09

  1. Yahoo!: Mike McKenna, Bene Franco
  1. Gender Formatting: (Gender Formatting Proposal)
  2. Add "inflection keywords" to CLDR
  3. make generic selector format
  4. name: selectformat?, default, switch, ..
  5. Markus to send sample of Google's plural patterns -> done: sent to icu-design list today, 2009-sep-09
  6. Mike to produce revised document.
  1. Canonicalizing ULocale (Mark)
  1. http://unicode.org/cldr/utility/languageid.jsp
  2. ZH-YUE-HANT_840
  3. http://unicode.org/cldr/utility/languageid.jsp?a=ZH-YUE%3F-HANT_840
  4. zh-abc-hant
  1. language: invalid
  2. script: hant
  1. Mark to update canonicalization for BCP47 changes. Pull data from res file, which comes from CLDR. Make C/Java use same data.
  1. Incoming bugs
  2. John Huân Vũ’s Last Meeting
  1. Thesis Survey: http://www.surveymonkey.com/s.aspx?sm=nUK04IMum23w8R9y5_2fKdrQ_3d_3d
  1. Cross script inheritance - follow-up from CLDR meeting ( John )
  1. From CLDR minutes:
  1. No script locale inherits from parent; language locale aliases to the default content script locale. Push all data down into scripts locales (one time).
  2. Only the default content script locale inherits from parent. Push non-default content into script locales (one time) (zh-Hant -> root, but zh-Hans -> zh -> root) -- run by ICU. If ok by them, then do it.
  3. Have marker in locale data that says that it doesn't inherit. (a) per item (b) overall (c) say what the parent is.
  1. In ICU:
  1. Use the marker Boolean %%ParentIsRoot:int{1}.
  2. This flag would be generated by LDML2ICUConverter from data already in CLDR.

2009/09/02

  1. ICU 4.3.2 schedule update? (Steven)
  1. 2009.10.07 - 5 weeks
  2. 4.3.2 milestone bugs
  3. Unicode schedule: not much time between UTC and Unicode 5.2 release (originally 9/18, now 9/30) (Markus).
  1. Conclusion: Unicode [ticket #7084] will not target M2 (4.3.2) but M3 (4.3.3) (API slush)
  2. concern over UCA/collation
  3. Unicode changelog will be updated. (source/data/unidata/changes.txt)
  1. Date format pattern support (Yoshito)
  1. Alternative patterns like "yyyy-MM-dd" vs "yyyy.MM.dd", "yyyy/MM/dd"
  1. see CldrBug:155 CldrBug:858 ? (srl)
  2. Could have alternates in CLDR data, and have ICU take it in?
  1. DateFormatPatternGenerator design question
  1. For skeleton pattern "yyMMdd", most locales return "default short format" or similar pattern
  2. However, in some locales, such as zh, it returns "yy年MM月dd日".
  1. IS_LEAP_MONTH in Calendar or ChineseCalendar (Jason)
  1. ICU4J has ChineseCalendar - which adds IS_LEAP_MONTH field
  2. ICU4J ChineseCalendar was ported to ICU4C.  At that time, IS_LEAP_MONTH was moved up to Calendar
  3. ICU4J has ChineseDateFormat, ChineseDateFormatSymbols.. ICU4C currently does not support ChineseCalendar formatting
  4. Question - #6033
  1. Should we move the logic/field from ChineseDateFormat/Symbols up to the super class at the same time?
  2. DateFormat.Field is synchronized with Calendar fields / ChineseDateFormat.Field has leap month field
  1. CLdr Bug # for further Chinese calendar discussion: (srl) CldrBug:1507
  1. Incoming bugs

2009/08/26

  1. Consider removing or shrinking unihan collation tailorings (radical-stroke sort orders) which are huge (1.95MB) (Markus & Mark)
  1. UCA 5.2: running short of primary-weight lead bytes unless we allocate weights more densely; could impact script reordering granularity (Markus & Mark)
  1. ICU4J 4.0.1.1
  1. Shipped the plug-in version to Eclipse
  2. GA today (2009-08-26) - updating the download page (draft - http://icu-project.org/download/draft4.0.html - will be removed after the official announcement.)
  1. Date format pattern support (Yoshito, next week)
  1. Alternative patterns like "yyyy-MM-dd" vs "yyyy.MM.dd", "yyyy-MM-dd"
  2. DateFormatPatternGenerator design question
  1. For skeleton pattern "yyMMdd", most locales return "default short format" or similar pattern
  2. However, in some locales, such as zh, it returns "yy年MM月dd日".
  1. Unicode 5.2 (Mark)
  1. Final data (sans UCA) final today
  2. Markus has a bug to integrate
  1. Incoming bugs

2009/08/19

  1. C vs. C++ APIs:  #3608 http://spreadsheets.google.com/ccc?key=tzB7XJEksLISbw8dTtFDS9g
  1. Public Link: http://spreadsheets.google.com/pub?key=tzB7XJEksLISbw8dTtFDS9g&output=html
  1. Ticket #7063 Change public/protected void methods to return this, for chaining (Yoshito / Mark)
  1. Change set http://bugs.icu-project.org/trac/changeset/26474 was temporarily reverted.
  2. Return type change requires client code changes if a user has a code extending an ICU class and override the method.
  3. Yoshito would like to keep JDK API compatibility
  4. change for internal, final, not-meant-to-be-subclassed.
  5. Mark to review changes that are not part of #3 and #4, and present proposal.
  1. ICU4J - Java compiler warning level (Yoshito)
  1. javadoc warnings are now enabled on Eclipse
  1. javdoc warning level question - Missing Javadoc tags: Ignored - at this moment
  2. Yoshito want to turn this on in future
  1. ant build to use -Xlint - #7097
  1. gender (yahoo)
  1. http://docs.google.com/Doc?docid=0Ady40I-34XO0ZDk3cXY5bl8xZ2Z3a3B4M3M&hl=en
  1. makefiles once again (srl - from support mailing list)
  2. Incoming bugs

2009/08/12

  1. ICU4J trunk breakage
  1. retracted Mark's chaining changes
  2. coverage test
  1. Incoming bugs
  2. ICU 4.4 planning - defer, not enough people in meeting and don't have tickets for all priority items
  3. Note: Peter on vacation Aug 17 - Sep 7, will try to have someone else from Apple call in

2009/08/05

  1. ICU 4.4 priorities planning, see http://sites.google.com/site/icusite/planning (Peter)
  1. http://sites.google.com/site/icusite/planning/icu-4-4
  2. Need to file bugs for tasks that do not have them, add bugs to each item in planning list
  1. Incoming bugs
  2. Note, due to UTC meeting, some people will not be able to participate next week (Mark)
  1. Someone else from Apple may be able to participate while Peter is on vacation (Aug 17 - Sep 7)
  1. CLDR issues: 08.01 1.8 Data submission starts

2009/07/29

  1. M1 status - Released 7/22/2009 w/ its own download page and mailing list announcement
  2. Unicode 5.2 scheduled for Sept 18.  Make milestone M2 happen 2-3 weeks after that?  CLDR also... (Markus, Mark)
  3. Chaining [Mark]- http://bugs.icu-project.org/trac/ticket/7063
  4. ICU Party [Mark] - 10 year party connected with conference dates, maybe first night of conference; past and present contributors. People don't care whether it is at Google or near the conference.
  5. Incoming bugs
  6. Public schedule [Mark] Mark to move above to public site. DONE
  7. STL [Mark]
  8. Reminder [Peter] - ICU 4.4 priorities planning next week, see http://sites.google.com/site/icusite/planning

2009/07/22

  1. M1 status (Michael, Yoshito) [4.3.1]
  1. C: Michael tagged C
  1. Markus: can include r26411? 
  2. A: michael is including r26414
  3. No errs in exhaustive test.
  1. J: build issues, working through them.
  1. Yoshito will update interim API docs
  1. We'll have a small M1 download page.
  1. can do API differences.
  1. Overall 4.4 planning - important enhancements, company priorities? (Peter)
  1. icusite / planning / 4.4 ?
  2. 2 weeks - Aug-5 meeting
  3. Markus to make a base page: http://sites.google.com/site/icusite/planning
  4. Peter to update/coordinate.
  1. Hardcode Unicode properties in ICU4J like in ICU4C? (Markus)
  1. Incoming bugs
  2. Items after bugs?

2009/07/15

  1. ICU in cars - raises issue, should we list products from non core members? (Mark)
  1. Mark to update page.
  1. ICU version number policy (Yoshito)
  1. X.X.Y.Z
  1. X.X -> Spec (API) version
  2. Y -> Maintenance release version for common changes (such as CLDR data updates, major functional updates in both C and J)
  3. Z -> C/J specific maintenance patch
  1. ICU4J 4.0.1 + small fixes for Eclipse -> 4.0.1.1
  1. Locale Display Names ( John ) http://bugs.icu-project.org/trac/ticket/4069
  1. Have CLDR data for localized display names, separators (we convert but don't use).
  2. Issue:  sorting. "British English", "US English" and "English" would sort away from each other.
  1. proposal: have existing API do what it does now, create new API for dialects.
  2. proposal: look at Mark's e-mail, look at alternatives..
  1. split format and parse?
  2. John to put together design doc.
  3. Issue: move LDN to i18n library in C? API stability? srl proposal
  1. move uloc_*getDisplayName* implementation into i18n
  2. make Locale::*getDisplayName* API turn into inlines that only work if linked against i18n
  1. String-replacement only Msgformat in common library?
  1. Little Red Switch for milestones? (Markus)
  1. add column of which BRS items are lRS - Yoshito and Markus
  1. M1 (4.3.1) - (Yoshito, Steven)
  1. code submission by july 17, ga july 22
  1. .res formatVersion 2 (Markus)
  1. Incoming bugs
  2. Document upgrading Unicode version: #6254 - Steven will send out e-mail with proposed contents.
  1. changelog in src/data/unidata/changes.txt

2009/07/08

  1. Confirmed in the meeting: Remove swapper support for resource bundle formatVersion 1.0 [pre ICU 2.8 (2003-11-22)]; http://bugs.icu-project.org/trac/ticket/7035 (Markus)
  2. ICU4J Collator dependencies removed from other core classes (Doug, Yoshito)
  1. Separated projects (icu4j-collate / icu4j-collate-tests)
  2. Other core code can now run without collator (icu4j-core-tests has no dependencies on icu4j-collate)
  3. breaking change for setLenientParse behavior - Doug to file bug to bring this back with reflection.
  4. Other low hanging fruit -
  1. transliteration? (sits on top of other pieces)
  2. formatting? (would need data file split)
  1. separate 'calendar' into core (FDOW, etc)
  1. timezone?
  2. locale ID must stay in core.. probably tied to LDN
  1. ICU4J Eclipse Java compiler warnings (Yoshito, John Vu)
  1. DEAD code such as if (false) {}
  2. Unused method/members (John Vu)
  3. http://icu-project.org/~johnvu/DeadCodeProposals.txt
  4. http://docs.google.com/View?id=dcsqkzbv_31dkq422dh
  1. ICU4C resource bundle formatVersion 2 (Markus)
  1. Incoming bugs
  2. Eclipse setup
  1. http://site.icu-project.org/setup/eclipse (needs update)
  2. http://site.icu-project.org/design/icu4j-development-with-eclipse
  3. http://site.icu-project.org/setup/subversion
  4. (also http://cldr.unicode.org/development/eclipse-setup)
  1. 4.3.1 (4.4M1) schedule
  1. Code submissions by July 17, 2009
  2. Once we have clean builds, tag 4.3.1 and move on

2009/07/01

  1. 4.2.1 status (Yoshito)
  1. Tagged release-4-2-1 for J - no changes from RC1
  2. Change set proposed by Jungshik is not yet merged - http://bugs.icu-project.org/trac/changeset/26196
  1. ICU4J clean up (Yoshito / Doug)
  1. Package com.ibm.icu.dev.test.util
  1. There are several types of classes
  1. Test cases for ICU4J API classes in com.ibm.icu.util
  2. Shared test utility code
  3. Test cases for 2. Shared utility code above
  1. Shared test utility code, such as ArrayCompartor, BagFormatter, BNF... are either never used by ICU4J API tests, or very limited references.  CLDR tooling heavily depending on these classes.
  1. Remove com.ibm.icu.dev.test.collator.RandomCollator
  1. timebomb was set to the test case on the same day when this was added to the repository
  2. http://bugs.icu-project.org/trac/ticket/5747
  3. It looks it had never worked.
  4. Dragging a some other dependencies
  1. Incoming bugs

2009/06/24

  1. Locale quality (message from Mark "Locale quality; for discussion in the core meeting" 6/23/09 1:26 PM)
  1. http://unicode.org/cldr/utility/languageid.jsp
  2. http://spreadsheets.google.com/pub?key=rLxL9P8USP0HtFLakzp_A_Q&output=html
  3. There are different ways to define 'what is a usable locale': formatting, etc
  4. Maybe split locale display data out into a separate tree? Maybe a separate getAvailable for names? May want to reduce what's put into ICU at LDML2ICUConverter time?
  5. Mark to run more numbers, compare to coverage in CLDR.
  6. May want to have run-time configuration parameter & metadata on locales: "i want locales with this level of completeness available"
  1. Tex: Load time configuration would be sufficient and allows users to minimize memory footprint.
  1. Possibly mark the CLDR data with a coverage level, to allow filtering.
  1. 4.2.1 status (John)
  1. 1.7.1 is in
  2. http://bugs.icu-project.org/trac/milestone/4.2.1
  1. all tickets: All Tickets 4.2.1
  1. several fixes/patches for the recent tzdata updates (7008, 7014, 7017)
  2. will merge changes into maint-4-2 and tag 4.2.1 rc1
  1. srl's work area for maint-4-2 C: (√: merge|commit|√)


  1. #
  1. r
  1. merged, committed, OK?
  1. ?
  1. -
  1. r26153
  1. m c r26160
  1. not targetted at 4.2.1, but a helpful testcase
  1. #6919
  1. r26063
  1. m c r26160
  1. icu-config escape
  1. #6979
  1. r26138 r26140 r26141
  1. m c r26160
  1. cldr 1.7.1
  1. #7020
  1. n/a - already in maint-4-2
  1. svn props
  1. #6918
  1. n/a - J
  1. Date formatting @numbers
  1. #6980
  1. r26055 r26056
  1. m c r26160
  1. smpdtfmt.h compiler warning
  1. #6908
  1. r25978
  1. m c r26160
  1. malayalam (!)
  1. release date is July1
  1. Proposed new collation features (Markus)
  1. Compact syntax for same-strength differences #7015
    CLDR <pc>abcduvwx</pc> -> ICU <*abcduvwx or maybe <*a-du-x
  2. Script reordering: [scriptReorder Grek Latn Zzzz Zyyy]
  3. Import: &a<b[import de-u-co-phonebk]&c<d[import und-Hani]
    plus a new attribute in the resource bundle like Private:int{1} for when we don't want the collator binary data to be built (e.g., for shared Unihan rules suffix)
  1. See CLDR design doc for script reordering and import: http://sites.google.com/site/cldr/development/design-proposals/collation-additions
  2. Q: Timeframe? A: compact syntax may be on a branch this weekend, only designs on others. So, by M2. No plans to change trunk for M1
  3. Design doc: http://sites.google.com/site/icusite/design/collation
  1. New member of ICU team at IBM (John Vũ)
  2. Incoming bugs
  1. #7007 is another DateTimePatternGenerator issue in ICU4C. The DateTimePatternGenerator has diverged quite a bit; Mark has made ICU4J changes after the original port to C, and Peter has made ICU4C fixes that need to be ported to J (see #6903). Peter to start e-mail discussion or document involving Peter, Mark, Claire about making the ICU4J and ICU4C implementations match, and about resolving how to address differences in behavior.

2009/06/17

  1. (fewer/no googlers?)
  2. IUC 33 update  (Steven)
  1. "What's new in ICU?" (Steven and Markus)
  1. icu4j relocation process - completed (Yoshito)
  1. Source layout changed for checkout
  1. change readmes,
  2. maybe http://site.icu-project.org/repository
  3. and the 'using eclipse with icu4j' page (already out of date)
  4. Yoshito has a ticket for updating the document.
  1. Build server/etc
  1. unchanged behavior, still have top level build.xml
  2. internal build server still working except for securetest on some platforms.
  1. Incoming bugs
  2. ICU4C lazy loading of tz strings, ticket #6879 (Peter); ICU4C uses const ZoneStringFormat object
  1. SimpleDateFormat & DateFormatSymbols have const methods that would cause lazy loading in the ZoneStringFormat object they reference (currently const)
  2. const in those methods is just semantic & can be cast away in this case; just make sure that object can be used from multiple threads (double-check locking)
  3. ensure user guide clairifies this usage of const (Peter to file ticket)

2009/06/10

  1. Resource bundle file format 2 (Markus)
  1. ICU4J proposed new source layout (Yoshito)
  1. See my message in icu-core ML posted yesterda
  1. Yoshito will wait for the feedbacks by 6/12
  2. If no major issues/objections, Yoshito will make the change in the trunk
  1. Additional proposals - agreed
  1. Move richedit (com.ibm.richtext) to icuapps repository
  2. Move tzu (TimeZone update utility) to tools repository
  1. Other notes
  1. Build target "fulldocs" - do we need this? -> no
  2. Skipping build target "indices" - modify ICU code to read res_index.res when res_index.txt is not found -> agreed
  3. "translitIMEJar" "indicIMEJar" - do we want to keep these build targets? -> keep them
  4. "swatDeprecated" "restoreDeprecated" - are these still important? -> keep them, consider annotation later
  1. Schedule
  1. 4.2.1- July 1st?
  1. coordinate with cldr 1.7.1
  1. 4.4M1 (4.3.1) July 22?
  1. java migration
  2. change in rb format
  3. java 5
  1. Please change the milestone to current as you check in changes.
  1. Products page updated with Apple section
  2. Incoming bugs

2009/06/03

  1. Making UnicodeSet and UnicodeMap inherit from Set<String>, SortedMap<String,T> after Yoshito's refactoring swat is done. (Mark)
  1. Mark to file bug. Must make sure all of the old API still works.
  2. Make sure old API still works
  1. Release things
  1. C 4.2.0.1 released (srl)
  2. CLDR 1.7.1 / 1.8 status (John, Mark)
  1. 171 - small but ugly bugs. slated for end of June.  ICU next milestone (421) would pick it up
  1. Review  (4.2.1 items) (srl)
  1. Old tickets (yoshito)
  1. http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&group=revw&milestone=4.1.2&milestone=4.2&milestone=4.1.4&order=priority
  1. ICU4J refactoring status (yoshito)
  1. source layout change
  1. Java Logger (michael)
  2. IUC#33 status?
  1. check back in a couple of weeks
  1. Incoming bugs

2009/05/27

  1. 4.2 API issues (Markus)
  1. 4.2 update release (srl)
  1. 4.2.0.1 (C) status  (srl)
  1. readme: http://source.icu-project.org/repos/icu/icu/tags/release-4-2-0-1/readme.html#News
  2. srl to post for reviewdone , GA tomorrow
  1. 4.0.2 (J) - (Yoshito) - collation performance regression
  2. 4.2.1 status
  1. when?
  1. will it pick up CLDR 1.7.1 data? (available in the next few weeks?)  Pr
  2. Probably before July 1
  1. Content?
  1. (4.2.1 items)
  2. Please review this list.
  3. John to add a ticket targetted for 4.2.1 
  1. Process
  1. Check in fixes to trunk
  2. they will be merged later into 4.2-maint
  1. Revisit list next week?
  1. srl to clean up 4.2.0.1 items off this list. DONE
  1. ICU4J refactoring (Yoshito)
  1. Java 5 generics support in core code (charset, impl, lang, math, text, util) are ready in Yoshito's branch
  1. Fixed ~1700 unchecked type warnings
  2. @SuppressWarnings("unchecked") for several instances
  1. generic array creation - eg. WeakReference<T>[]
  2. Casting value returned by clone()
  1. Comparable/Comparator with generics
  1. Removed compareTo(Object),  compare(Object, Object)
  2. Added (or preserved) compareTo(T), compare(T, T)
  1. no plan to spend time to work on non-core codes (tests, tools, demo) - ~2700 warnings in these codes
  2. Yoshito wants to separate the build unit for core code from others
  1. Independent source root directory / Eclipse project files
  1. Next steps - modularization of core code and data
  1. Had a brainstorm session this morning
  2. Attendees agreed -
  1. Break ICU data into smaller pieaces - aligned to each function
  2. Separate project per function and reduce inter-service dependencies
  1. Markus' new intern Michael Fairley
  1. IcuWiki (http://source.icu-project.org/cgi-bin/icuwiki.pl): is it used? do or should we advertise it? do or should we monitor what users put there? (Markus)
  1. Not used in 2.5 years
  2. Open wiki could attract spam
  3. Better cull information/requests/FAQs from the public mailing lists (icu-support & icu-design)
  4. If we wanted another forum, we could start a blog and look for comments
  5. Mark: Can embed a Google gadget on a Sites page; e.g., a blog gadget
  6. Any public forum needs to be moderated
  7. Steven to make a pass through IcuWiki, make sure contents is elsewhere, and shut it down: ticket #6974
  8. Mark: We could edit the ICU Wikipedia page
  1. IUC33 - deadline is tomorrow May 28th (srl)
  1. proposed items
  1. Incoming bugs
  1. defer to next time.

2009/05/20

  1. 4.2 update release (srl)
  1. Everything is noted on the downlo
  2. please finish reviews
  3. #6953: post 4.2 patches
  4. Target GA date : May 25th (ICU4C 4.2.0.1)
  1. Closing 4.2 tickets

4.2 tickets

  1. CLDR schedule (John)
  1. Latest is at http://cldr.unicode.org/
  1. ICU4J clean up / migrating to Java 5+
  1. Removing conditional compilation blocks for old Java versions -> #6952
  2. Generics support -> #6951
  3. Please don't damage icu4j trunk for 1 more week.
  1. Incoming bugs

2009/05/13

  1. 4.2 released
  1. trunk status
  1. Known bugs and 4.2.0.1 candidates
  1. 4.2.1  ? 4.2.0.1 ? (srl)
  1. Could ship patch., and catch up in 4.2.1
  2. Or, could ship patches as 4.2.0.1 for C
  3. Today:  
  1. srl: update errata on download page
  2. srl: post notice to icu-support, icu-announce.
  1. Monday: deadline for 4.2.0.1 list 
  1. please update the agenda item #1 above and notifu
  1. RC:  5/20
  2. GA: 5/25
  1. UnicodeString vs. valgrind/Purify (Markus)
  1. ICU data file size reduction (Markus)
  1. Java 5 syntax migration(yoshito)
  1. Anyone working on ICU4J trunk actively now?
  1. hold off on changes to C/J trunk for another week..
  1. generics
  1. Incoming bugs

2009/05/06

  1. 4.2 status
  1. Latest ICU4C 4.2 build - d04 (on download page)
  2. Latest ICU4J 4.2 build - d03
  3. ICU4J - no known issues, no plans for code change. d03 is RC
  4. ICU4C build on IBM iSeries
  1. makefiles: autoconf broken, will have errata on download page
  2. C linkage: similar to r25954 (will not affect other platforms).  Will revisit U_CALLCONV in 4.4.
  1. plan: 1 more build for C, update files and release early friday
  1. NOTICE file or similar: List files with non-ICU licenses (Markus)
  1. pull from ibm's scan
  1. ICU data file version numbers? (Markus)
  1. today's e-mail to icu-core
  1. CLDR data comparison in ICU download page (Mark)
  1. Mark to update or notify icu-core when CLDR 1.7 language is complete.
  1. Incoming bugs
  1. Note: this is an updated "incoming bugs" link. I've now updated the "incoming bugs" link on the trac main page. If you have a bookmark to incoming, you may want to change it.  -srl

2009/04/29

  1. 4.2 status update
  1. maint-4-2 branches were created / d01 tagged for both C and J
  2. Complete code review!!! http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&group=revw&order=priority&revw=!
  3. Open issues
  1. ICU4C build issues
  1. http://bugs.icu-project.org/trac/ticket/6888 - conversion failure (optimizer error) on ICC 9.0
  1. will document as known issue.
  1. http://bugs.icu-project.org/trac/ticket/6890 - StringPiece::npos duplicate - msvc9 64-bit
  1. - resolved w/ additional #ifdef.
  1. http://bugs.icu-project.org/trac/ticket/6894 - numerics failure on ICC 11.0
  1. will document as known issue
  1. Z/os issue? (on mailing list)
  1. UnicodeString&*
  1. RBNF / stack overflow problem
  1. Several updates in maint-4-2 after d01 tag
  1. http://bugs.icu-project.org/trac/ticket/6892 (C and J) - CLDR - Integrate backout of CLDR#2045
  2. http://bugs.icu-project.org/trac/ticket/6891 (J only) - Remove foundation jar file from the ICU4J source package
  3. http://bugs.icu-project.org/trac/ticket/6893 (J only) - LMBCS inconsistency between ICU4C and ICU4J
  4. r25946 (C) readme update
  1. Any other issues to be addressed for 4.2 release?
  2. Trunk: keep post-4.2 changes off for now, use a branch until the release is GA..
  3. All: Review release notes (readmes) & download page.
  1. ICU info XML updated for C   http://source.icu-project.org/xml/icumeta.xml
  1. BRS items
  1. Incoming bugs

Usage of user guide:

Pageviews

Unique Views

Bounce Rate

Top Content

Pages

Pageviews

% Pageviews

/

2,510

13.67%

/strings/regexp

1,054

5.74%

/icufaq

851

4.63%

/strings

812

4.42%

/conversion

566

3.08%

2009/04/22

  1. Check status for code freeze {yoshito, srl}
  1. delay checkin deadline to apr27 ( Markus )
  2. Yoshito to make a spreadsheet of things that can't make it by the end of the day.
  3. Update spreadsheet and email core when there is a change.  Don't check in changes not targetted for 4.2 after wednesday if not on the spreadsheet.
  4. The spreadsheet is here -> https://spreadsheets.google.com/ccc?key=plRXpLeUpVd1gHyokul9VYw&hl=en
  1. (C) assign <srl>
  2. BRS items
  3. Incoming bugs
  4. Lenient number/date/time parsing (Eric)
  5. Secret Agenda item [Mark] ~10m - http://docs.google.com/Doc?id=dfqr8rd5_138xsbgchq

2009/04/15

  1. ICU 4.2 / CLDR 1.7 updates
  1. version number to 4.2 today?
  2. upcoming CLDR 1.7 data changes
  3. need upcoming milestones in Trac  ? (Markus) -- no, just use 4.2
  4. CF schedule - Yoshito's proposal below
  1. Any changes other than CLDR related stuff should be in by the end of April 22 (next Wed)
  2. We'll check the status in the next PMC meeting
  3. For any changes after April 22, the developer must communicate with Steven/Yoshito and also post a note to core ML
  4. Will start full testing / work on release files on April 23
  1. BRS items
  1. ticket 6799 - C warnings - many removed (srl), please do a build and help with remaining ones
  1. ticket 6775 - collation key difference between C and J
  1. Claire probably does not have time to look into this before 4.2
  2. Yoshito to take over the ticket?
  1. Incoming bugs
  1. The following is really also incoming but got accidentally assigned: #6846

2009/04/08

  1. Google Meeting Conflict, most of us will probably miss today.  (andy)
  2. CLDR status update (John,  Jason running in parallel)
  1. looking good..  probably something by tomorrow or friday. 
  2. Peter can run tests as well.
  1. 4.2M5 - 2009/04/13th
  1. After end of Thursday (2009/04/09): Communicate to core mailing list for any non-CLDR items.
  2. (Michael to take lead on C release for M5)
  3. C: Autoconf 2.63b  - beta [announcement] . Not going to switch to this for M5. May have a branch for testing later (srl).
  4. Reviews!
  1. Calendar::createInstance to return U_USING_DEFAULT_WARNING (Yoshito)
  1. http://bugs.icu-project.org/trac/ticket/6827
  2. post 4.2 - discuss next week?
  1. Layout Engine Testing, fonts availability. (andy)
  1. were sent to Andy.
  2. bug #6829 - andy to look at, work w/ john
  1. Adding private functions (Peter)
  1. new methods in reldtfmt.h, new udat_impl.h?
  2. put in public but @internal (technology preview)
  1.  * @internal ICU 4.0.1 technology preview
  1. bugs
  1. (some deferred until next time)
  1. BRS
  1. Please look at BRS items (put links in top of agenda)

2009/04/01

  1. 4.2 BRS items (Yoshito)
  1. http://spreadsheets.google.com/ccc?key=plRXpLeUpVd0Zk0M9CEFCLg
  2. BRS task description - http://site.icu-project.org/processes/release
  1. CLDR status update
  1. CLDR BRS - http://sites.google.com/site/cldr/development/cldr-big-red-switch
  2. M5 4.1.5 April 13, 2009 -> wait until CLDR 1.7 RC is ready to go
  3. CLDR 1.7 d01 tagged version going into ICU 4.2 M5 (around April 20?)
  1. Territory based locale service (Yoshito)
  1. http://bugs.icu-project.org/trac/ticket/6816
  2. ICU Locale Service framework depends on availability of locale data (requiring a language)
  1. Locale capacity constants? (Yoshito)
  1. http://icu-project.org/apiref/icu4c/uloc_8h.html
  2. #define ULOC_FULLNAME_CAPACITY   56
  3. #define ULOC_KEYWORDS_CAPACITY   50 and #define ULOC_KEYWORD_AND_VALUES_CAPACITY   100
  1. "Useful constant for the maximum size of keywords in a locale." vs. "Useful constant for the maximum SIZE of keywords in a locale."???
  1. tz 2009d resolution?
  2. bugs
  3. ICU site transition:  gopher://gopher.icu-project.org:70/ - TEH AWSUM
  1. Markus: I get "Access Denied"
  2. Doug: Works for me.  I look forward to seeing more APIs implemented...
  3. srl: doesn't work in Chrome/win. Works for me in Win Firefox 3.0.7

2009/03/25

  1. CLDR 1.7 status (John)
  1. next to last integration
  2. icu m 4.1.4 has a snapshot 
  3. CLDR deadline extended a week. ST closes next tuesday Mar31. 
  4. (proposal)
  1. Apr 8- CLDR RC, John begins testing.
  2. M5 Apr 13 - waiting for CLDR RC
  3. +10 days M6 code freeze- Apr 27
  4. GA: try for May 8
  1. Metazone Argentina_Western problem (yoshito)
  1. For last half year, there were no zones mapped to the metazone Argentina_Western
  2. In tzdata 2009d, America/Argentina/San_Luis was updated and Argentina_Western is the most appropriate mapping
  3. The current CLDR golden zone (America/Argentina/Mendoza) for Argentina_Western no longer mapped to the metazone
  4. golden zone mappings are in supplementalData.res, which has never been shipped as a part of tz update
  5. There is no appropriate golden zone which satisfies the criteria
  6. It looks there is no single solution other than ignoring the metazone working for 3.8.x and 4.0.x
  1. UCharacter.isXXXX in ICU4J (yoshito) - http://bugs.icu-project.org/trac/ticket/6689
  1. J: UCharacter.isWhitespace
  1. U+2007 (FIGURE SPACE) -> JDK:false / ICU:true
  2. U+200B (ZERO WIDTH SPACE) -> JDK:true / ICU:false
  3. JDK JavaDoc -> Determines if the specified character (Unicode code point) is white space according to Java. A character is a Java whitespace character if and only if it satisfies one of the following criteria: - It is a Unicode space character (SPACE_SEPARATOR, LINE_SEPARATOR, or PARAGRAPH_SEPARATOR) but is not also a non-breaking space ('\u00A0', '\u2007', '\u202F').
  4. Character category for U+200B was updated in Unicode 4.1.  Java 6 still uses Unicode 4.0, while ICU uses Unicode 5.1.
  5. ICU has UCharacter.isUWhitespace for supporting the Unicode whitespace definition
  6. If UCharacter.isWhitespace is just for JDK compatibility support, why do we need our own implementation?
  1. UCharacter.isJavaIdentifierStart/isJavaIdentifierPart simply call the JDK method (with downcasting from int to char!!!)
  1. ICU4J:  isWhitespace uses ICU Unicode data (5.1)
  1. Proposal:
  1. ICU4C and ICU4J should return same values for corresponding API
  2. Routines specified as "returning what Java returns" (isJavaSpaceChar?) should return a value coordinated with the latest shipping Java. (document this)
  1. These aren't shadows (they do not call through to JDK).
  1. Need a list of all APIs and differences of behavior.
  1. Yoshito and Steven to collect APIs.
  1. isDigit - is this in the same category?
  1. API reference pages on icu-project site (yoshito)
  1. Two versions - one for the last public release version, one for the latest milestone.
  1. Release tasks(Steven)
  1. copyright assignment: Google since 10/2007 (Last: Jungshik and Jennifer), others?
  1. srl to talk to Markus next week
  1. Yoshito to create BRS worksheet for 4.2, will discuss about task assignment in next PMC (April 1)
  1. bugs
  1. All bugs in C trunk since 4.0 was branched(login first)
  2. All tests in J trunk since 4.0 was branched(login first)

2009/03/18

  1. M4 status
  1. C: API docs not updated ( @status )
  1. builds clean inc. exhaustive
  2. srl to finish this afternoon-hold off on C checkins, srl to update core list
  3. http://icu-project.org/~srl/api-tmp/APIChangeReport-401vsr25617.html
  1. J: M414 already tagged, API docs posted. Done.
  2. will set version "4.1.5" post M4
  3. After M5 (Apr 8 ) will be "4.2"
  1. C: doc format for bytestream/ stringpiece, etc.
  1. may be first public API not originally for ICU.
  2. //@draft -> ///@draft
  1. uenum_openStringEnumeration() vs. uenum_openFromStringEnumeration()   (Andy & Markus)
  1. do this, for M4
  1. "Philosophical" question about the specificity of request patterns for date-time pattern generation (Peter). Perhaps this is a question for the CLDR TC instead, but the behavior is closely tied both to the CLDR data and to the ICU code that uses it.
  1. Q: Warnings?
  1. deprecated calls - getLocale
  1. rewrite or #ifdef out?
  1. signed/unsigned
  2. priority: library code, then test
  1. get warnings from includes of public ICU header files  out first
  1. Q: should strict be default for the end user?
  2. srl has #6799, Will split up the work.
  3. need to test warnings on multiple platforms.
  4. Need to have a 'cookbook' of how to fix typical warnings.
  5. M5: no warnings. Re-evaluate if too hard.
  1. other

2009/03/11

  1. M4 status
  1. Final check in for M4: March 13 (Friday)
  1. Impact of CLDR 1.7 schedule (John/Mark?)
  1. Move schedule by a week depending on CLDR
  2. Markus: Propose move M5, M6, and GA by a week (approved by all)
  1. Locale syntax (Yoshito) (Locale Enhancement Project )
  1. BCP47 extensions (other than LDML keywords)/private use to retro fit to the existing keyword syntax such as en_US@a=foo-bar;x=yoshito-umaoka for BCP47 tag en-us-a-foo-bar-x-yoshito-umaoka
  2. Do not change existing Locale keyword APIs - that means, all @key=value are interpreted as keywords including non-LDML extensions/private use
  3. We want to keep the syntax/method synchronized with the proposal for JDK7.  That means, new API set has a distinction between LDML keywords and other extensions.
  4. en-us-u-co-phonebk-ca-japanese  ->  en_US@collation=phonebook;calendar=japanese
  5. 4.2 support:
  1. toLanguageTag [ Locale -> BCP47 ],
    fromLanguageTag [BCP47 -> Locale]  - both map  ca<->calendar, phonebk<->phonebook,  map single letter extensions and private use to/from single letter ICU locale key/values.
  1. Dave has a document on support for bcp47 such as de__PHONEBOOK.. issues were created for all things found.
  1. it's in trac. yoshito to look at.
    Dave to send around. Done.
  1. For now, if you want to use BCP47 with ICU, convert it when it comes in into icu format.
  1. Yoshito to file ticket to evaluate this in 4.4
  1. Bugs

2009/03/04

  1. Google analytics
  1. 4.2 API change deadline
  1. Steven/Yoshito will generate API change report for review on March 11.
  2. Please communicate with team using the core ML for any public API signature changes/additions beyond March 10
  3. Final check in for M4 - end of March 13.
  1. Clean up past milestones!
  1. Retarget tickets missed the original target milestone
  2. Finish code review for 4.1.1/4.1.2/4.1.3
  1. Bugs
  2. Java Locale meeting - JDK 7.0
  1. http://sites.google.com/site/openjdklocale/

2009/02/25

  1. Locale identifier discussion update (Yoshito)
  1. LDML to use BCP47 syntax for locale identifiers
  2. Some key/type names will have short canonical name
  3. Long/short key/type name mappings will be in supplemental data in 1.7
  4. No type name changes in locale data files for now.  (In other words, we do not change "type=phonebook" to "type=phonebk" in locale data file)
  5. ICU 4.2 to provide conversion function between locale and BCP47 language tag
  1. Bugs
  2. Site changing to Sites - update http://site.icu-project.org / [Mark, Markus]
  1. Redirected pages - http://site.icu-project.org/charts/comparison - Michael
  2. Lock old pages - Michael
  3. Would be good to add Products from Core members: Google, Apple, NetApp; plus others on request.
  4. Should have announcement about change.

2009/02/18

  1. M3 status / Next milestone
  1. Tagged M3 for both C and J (4.1.3)
  2. M4 (API Freeze) (4.1.4) March 15 -> 18
  1. please move the target if you won't make it
  1. M5 - CLDR Final Candidate Data - April 1, 2009.
  2. Code review -
  1. 4.1.1 -> http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&group=revw&milestone=4.1.1&order=priority&revw=!
  2. 4.1.2 -> http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&group=revw&milestone=4.1.2&order=priority&revw=!
  3. 4.1.3 -> http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&group=revw&milestone=4.1.3&order=priority&revw=!
  1. ICU User Guide: Site contents complete (Markus)
  1. ICU main site: Site contents complete (Mark)
  1. UnicodeString <-> UTF-32/UTF-8/string API (Markus, ticket 6677)
  1. http://www.macchiato.com/unicode/nfc-faq
  2. Farewell to Eric
  3. Bu/gs!

2009/02/11

  1. 4.2M3 - API slush on Feb 18 (Steven, Yoshito)
  1. Any "major" API changes not yet done?
  2. Checkin deadline Feb 13, 6pm PT
  3. If you cannot make it, please post a message to core ML
  1. FYI: CLDR 1.8 tentative schedule
  1. Release date - March 10, 2010
  2. Final Candidate - Feb 10, 2010
  3. Data Resolution Start - Jan 13, 2010
  4. Data Vetting Start - Nov 24, 2009
  1. ICU 4.4 tentative major milestone schedule proposal
  1. API slush - Nov 4, 2009
  2. API freeze - Jan 13, 2010
  3. Code freeze - Feb 24, 2010
  4. Early milestones
  1. July 2009 (M1)
  2. September 2009 (M2)
  1. Updates in supplemental data from CLDR might be integrated in each ICU milestone
  2. Unicode 5.2 - Sep-Nov.  We may need an extra milestone if we cannot make it before API slush.
  1. 6317- api question (for Andy)
  1. legacy grapheme clusters vs extended grapheme clusters
  2. whether we need to support legacy grapheme clusters (currently, ICU supports extended only)
  3. Andy will reply to the bug
  4. Update the user guide
  1. ICU APIs support CLDR data (Claire)
  1. Motivation - We use ICU for testing CLDR data.  We want to make sure that every CLDR data is actually used by ICU, therefore, you can reasonably test CLDR data via ICU.
  1. Support of STL string class (Markus)
  1. Any interests from companies other than google?
  2. Apple uses C API
  3. IBM - currently not
  1. Bugs!

2009/02/04

  1. Several people will be absent due to Unicode UTC meeting
  2. API slush - checkins by Friday(13) next week - get proposals to the design list by this friday (6th)
  3. Yoshito: proposal: shorten key/value in CLDR identifier, may impact ICU. ICU may need to support truncated form of locale IDs.
  4. language tags ↔ locale ids - error mode, what to do in invalid case
  1. shouldn't happen in normal case
  2. could have a 'validate' function - determine if everything is of valid length and structure.
  3. expect try/catch around every use of a language tag?
  4. Distinguish 'fail to parse' from 'parse, but unrecognized'.  ( invalid vs. unknown )
  1. Bugs!

2009/01/28

CLDR status- in progress, probably need to check on new features in 1.7 to make sure they get implemented on ICU. Will file umbrella "make sure everything is in sync" bug. (John)

4.1.2 (Steven, Yoshito)

  1. completed both C and J
  2. version number to 4.1.3 (J done / C - srl to do. ?)
  3. r/t transliterator test bumped,
  1. will disable those in C and J. Already a CLDR bug [srl to find #] to implement these in CLDR.
  1.  and dt pattern generator bumped -Yoshito sent out a note and will create ticket.
  2. code review! will finish review of 4.0 bugs, please work on later ones also.
  3. Next milestone - (4.1.3) M3 Feb 18 / API slush - large API changes/addition must be in by M3
  4. Fri Feb 13, API checkin slush deadline. 

4.4 plan (Yoshito)

  1. need to sync with CLDR 1.8
  1. CLDR could always have minor releases that update to newest Unicode (already do
  1. sync with Unicode 5.2? October 2009. Will want a milestone to include that version.  (6.0 slated for fall 2010)
  2. Annual major release? Apr-May timeframe,
    then 1-2 maintenance releases between?
  3. More frequent milestone?
  4. March 10, 2010 - proposed 4.4 GA / propose the date to CLDR project next week

User Guide: Go forward with move to Site version (also at http://userguide.icu-project.org/ -- needs links fixed), or withdraw?

  1. go forward!
  2. Mark: didn't work on icu site more, maybe in some weeks.

Bugs

2009/01/21

4.0.1 status - shipped!

4.1.2 progress

  1. J - fine
  2. C -
  1. exhaustive mode cintltst failures ( ucnvsel / propsvec ) (markus and michael working on)
  1. Update: other tests pass.
  1. noted: AIX/GCC fails to build in strict(bug filed)
  2. noted: VS 2003 optimized fails bidi get mirrored char (srl to note in readme)
  3. noted: srl will start weekly exhaustive build
  4. 4.1.2 was tagged.. when there's a fix for ucnvsel, will merge it in to the tag and retest.

timezone 2009a (Yoshito)

  1. 'handle exception' in tz2icu on Cygwin.. looking ?
  2. changes to zic? take a look at any differences since tzcode was put into icu.
  3. Yoshito and John to look at?

Yoshito: will change CodeMangler to default to Java 6.

  1. eclipse IDE needs to be changed from workplace IDE 5 to Java 6. Yoshito to send instructions.
  2. Otherwise, strange and spurious checkin differences!
  3. there is an ant target that will reset the source code to prepare for checkin.
  4. Yoshito will checkin project file update update

Bugs?

  1. Bugs!

Please look at your bugs to review.

EOM (end of meeting..)

2009/01/14

ICU 4.0.1 status (Steven, Yoshito)

  1. ICU4C revision: r25245
  2. ICU4J revision: r25219 / tagged
  3. 4.0.1C candidate in testing
  4. should go out tomorrow (Thu, Jan 15)
  5. Beta download page up today.

ICU 4.1.2 schedule - M2 upcoming next Wed (1/21)

  1. #6433 xiaomei - do not need to pull out the current work-in-progress code in trunk - if xiaomei has time, merge current updated code from branch to trunk
  2. Friday noon PST - once stop code submission to trunk
  3. Any check-in after Friday noon - send a note to icu-core list
  4. Let M2 out as scheduled
  5. Add M3 (Feb 18) to catch up missing features

ICU 4.2 Platforms

  1. C/J: IBM binary/testing plans (Steven, Yoshito)
  2. C: VC 2005 -> 2008?
  3. IBM test:

  1. Windows Vista (32-bit), Visual Studio 2005 2008 express

    Windows Server 2003 2008 (64-bit), Visual Studio 2005 2008AIX 6.1 VisualAge 9 (64-bit)
    Solaris 10 (64-bit), Sun Studio 12
    Red Hat Enterprise Linux 5 (32-bit), GCC 4.1.2
    '''+Red Hat Enterprise Linux 5 (64-bit), GCC 4.1.2
    HP 11iv3 IA64 (64-bit) - aCC A.6.14
  2. ICU4J
    Note: Class files are Java 1.3 compatible (javac -target 1.3)
  1. Windows Vista (32-bit) - IBM JDK 6 (32-bit)
  2. Solaris 10 (SPARC 64-bit) - Sun JDK 6 (32-bit)
  3. AIX 6.1 - IBM JDK 6 (64-bit)
  1. Drop MSVC6 - no support

Brief discussion about MSVC 6:

Upgrade Java source file syntax level

Reminder - closing milestone 4.0 / 4.1.1

  1. We still need several tickets reviewed for 4.0

Bugs

2009/01/07

  1. ICU 4.0.1 status (Steven/Yoshito)
  1. Steven and Yoshito to update
  1. C++ static cast (Andy)
  1. Andy to try some in a branch - IBM team tries it with various compilers
  2. If it works, we will use them in new code.. Probably not trying to update everything at one time. 
  1. ICU Numbering System proposal (John)
  1. John to send out the proposal shortly.
  2. Force latin supported?
  1. The proposal is more comprehensive
  2. Yes? - Can be implemented like this way.
  3. How can we handle alternate decimal symbols (decimal point, grouping separator), not only for digits (0-9)?
  1. Bugs

************************

************************

2009/03/25

  1. CLDR 1.7 status
  2. Metazone Argentina_Western problem (yoshito)
  1. For last half year, there were no zones mapped to the metazone Argentina_Western
  2. In tzdata 2009d, America/Argentina/San_Luis was updated and Argentina_Western is the most appropriate mapping
  3. The current CLDR golden zone (America/Argentina/Mendoza) for Argentina_Western no longer mapped to the metazone
  4. golden zone mappings are in supplementalData.res, which has never been shipped as a part of tz update
  5. There is no appropriate golden zone which satisfies the criteria
  6. It looks there is no single solution other than ignoring the metazone working for 3.8.x and 4.0.x
  1. UCharacter.isXXXX in ICU4J (yoshito) - http://bugs.icu-project.org/trac/ticket/6689
  1. isWhitespace
  1. U+2007 (FIGURE SPACE) -> JDK:false / ICU:true
  2. U+200B (ZERO WIDTH SPACE) -> JDK:true / ICU:false
  3. JDK JavaDoc -> Determines if the specified character (Unicode code point) is white space according to Java. A character is a Java whitespace character if and only if it satisfies one of the following criteria: - It is a Unicode space character (SPACE_SEPARATOR, LINE_SEPARATOR, or PARAGRAPH_SEPARATOR) but is not also a non-breaking space ('\u00A0', '\u2007', '\u202F').
  4. Character category for U+200B was updated in Unicode 4.1.  Java 6 still uses Unicode 4.0, while ICU uses Unicode 5.1.
  5. ICU has UCharacter.isUWhitespace for supporting the Unicode whitespace definition
  6. If UCharacter.isWhitespace is just for JDK compatibility support, why do we need our own implementation?
  1. UCharacter.isJavaIdentifierStart/isJavaIdentifierPart simply call the JDK method (with downcasting from int to char!!!)
  1. isDigit - is this in the same category?
  1. API reference pages on icu-project site (yoshito)
  1. Two versions - one for the last public release version, one for the latest milestone.
  1. Release tasks - copyright assignment (Steven)
  1. Google since 10/2007 (Last: Jungshik and Jennifer), others?
  2. Yoshito to create BRS worksheet for 4.2, will discuss about task assignment in next PMC (April 1)
  1. bugs

2009/03/18

  1. M4 status
  1. C: API docs not updated ( @status )
  1. builds clean inc. exhaustive
  2. srl to finish this afternoon-hold off on C checkins, srl to update core list
  3. http://icu-project.org/~srl/api-tmp/APIChangeReport-401vsr25617.html
  1. J: M414 already tagged, API docs posted. Done.
  2. will set version "4.1.5" post M4
  3. After M5 (Apr 8 ) will be "4.2"
  1. C: doc format for bytestream/ stringpiece, etc.
  1. may be first public API not originally for ICU.
  2. //@draft -> ///@draft
  1. uenum_openStringEnumeration() vs. uenum_openFromStringEnumeration()   (Andy & Markus)
  1. do this, for M4
  1. "Philosophical" question about the specificity of request patterns for date-time pattern generation (Peter). Perhaps this is a question for the CLDR TC instead, but the behavior is closely tied both to the CLDR data and to the ICU code that uses it.
  1. Q: Warnings?
  1. deprecated calls - getLocale
  1. rewrite or #ifdef out?
  1. signed/unsigned
  2. priority: library code, then test
  1. get warnings from includes of public ICU header files  out first
  1. Q: should strict be default for the end user?
  2. srl has #6799, Will split up the work.
  3. need to test warnings on multiple platforms.
  4. Need to have a 'cookbook' of how to fix typical warnings.
  5. M5: no warnings. Re-evaluate if too hard.
  1. other

2009/03/11

  1. M4 status
  1. Final check in for M4: March 13 (Friday)
  1. Impact of CLDR 1.7 schedule (John/Mark?)
  1. Move schedule by a week depending on CLDR
  2. Markus: Propose move M5, M6, and GA by a week (approved by all)
  1. Locale syntax (Yoshito) (Locale Enhancement Project )
  1. BCP47 extensions (other than LDML keywords)/private use to retro fit to the existing keyword syntax such as en_US@a=foo-bar;x=yoshito-umaoka for BCP47 tag en-us-a-foo-bar-x-yoshito-umaoka
  2. Do not change existing Locale keyword APIs - that means, all @key=value are interpreted as keywords including non-LDML extensions/private use
  3. We want to keep the syntax/method synchronized with the proposal for JDK7.  That means, new API set has a distinction between LDML keywords and other extensions.
  4. en-us-u-co-phonebk-ca-japanese  ->  en_US@collation=phonebook;calendar=japanese
  5. 4.2 support:
  1. toLanguageTag [ Locale -> BCP47 ],
    fromLanguageTag [BCP47 -> Locale]  - both map  ca<->calendar, phonebk<->phonebook,  map single letter extensions and private use to/from single letter ICU locale key/values.
  1. Dave has a document on support for bcp47 such as de__PHONEBOOK.. issues were created for all things found.
  1. it's in trac. yoshito to look at.
    Dave to send around. Done.
  1. For now, if you want to use BCP47 with ICU, convert it when it comes in into icu format.
  1. Yoshito to file ticket to evaluate this in 4.4
  1. Bugs

2009/03/04

  1. Google analytics
  1. 4.2 API change deadline
  1. Steven/Yoshito will generate API change report for review on March 11.
  2. Please communicate with team using the core ML for any public API signature changes/additions beyond March 10
  3. Final check in for M4 - end of March 13.
  1. Clean up past milestones!
  1. Retarget tickets missed the original target milestone
  2. Finish code review for 4.1.1/4.1.2/4.1.3
  1. Bugs
  2. Java Locale meeting - JDK 7.0
  1. http://sites.google.com/site/openjdklocale/

2009/02/25

  1. Locale identifier discussion update (Yoshito)
  1. LDML to use BCP47 syntax for locale identifiers
  2. Some key/type names will have short canonical name
  3. Long/short key/type name mappings will be in supplemental data in 1.7
  4. No type name changes in locale data files for now.  (In other words, we do not change "type=phonebook" to "type=phonebk" in locale data file)
  5. ICU 4.2 to provide conversion function between locale and BCP47 language tag
  1. Bugs
  2. Site changing to Sites - update http://site.icu-project.org / [Mark, Markus]
  1. Redirected pages - http://site.icu-project.org/charts/comparison - Michael
  2. Lock old pages - Michael
  3. Would be good to add Products from Core members: Google, Apple, NetApp; plus others on request.
  4. Should have announcement about change.

2009/02/18

  1. M3 status / Next milestone
  1. Tagged M3 for both C and J (4.1.3)
  2. M4 (API Freeze) (4.1.4) March 15 -> 18
  1. please move the target if you won't make it
  1. M5 - CLDR Final Candidate Data - April 1, 2009.
  2. Code review -
  1. 4.1.1 -> http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&group=revw&milestone=4.1.1&order=priority&revw=!
  2. 4.1.2 -> http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&group=revw&milestone=4.1.2&order=priority&revw=!
  3. 4.1.3 -> http://bugs.icu-project.org/trac/query?status=new&status=assigned&status=reopened&group=revw&milestone=4.1.3&order=priority&revw=!
  1. ICU User Guide: Site contents complete (Markus)
  1. ICU main site: Site contents complete (Mark)
  1. UnicodeString <-> UTF-32/UTF-8/string API (Markus, ticket 6677)
  1. http://www.macchiato.com/unicode/nfc-faq
  2. Farewell to Eric
  3. Bu/gs!

2009/02/11

  1. 4.2M3 - API slush on Feb 18 (Steven, Yoshito)
  1. Any "major" API changes not yet done?
  2. Checkin deadline Feb 13, 6pm PT
  3. If you cannot make it, please post a message to core ML
  1. FYI: CLDR 1.8 tentative schedule
  1. Release date - March 10, 2010
  2. Final Candidate - Feb 10, 2010
  3. Data Resolution Start - Jan 13, 2010
  4. Data Vetting Start - Nov 24, 2009
  1. ICU 4.4 tentative major milestone schedule proposal
  1. API slush - Nov 4, 2009
  2. API freeze - Jan 13, 2010
  3. Code freeze - Feb 24, 2010
  4. Early milestones
  1. July 2009 (M1)
  2. September 2009 (M2)
  1. Updates in supplemental data from CLDR might be integrated in each ICU milestone
  2. Unicode 5.2 - Sep-Nov.  We may need an extra milestone if we cannot make it before API slush.
  1. 6317- api question (for Andy)
  1. legacy grapheme clusters vs extended grapheme clusters
  2. whether we need to support legacy grapheme clusters (currently, ICU supports extended only)
  3. Andy will reply to the bug
  4. Update the user guide
  1. ICU APIs support CLDR data (Claire)
  1. Motivation - We use ICU for testing CLDR data.  We want to make sure that every CLDR data is actually used by ICU, therefore, you can reasonably test CLDR data via ICU.
  1. Support of STL string class (Markus)
  1. Any interests from companies other than google?
  2. Apple uses C API
  3. IBM - currently not
  1. Bugs!

2009/02/04

  1. Several people will be absent due to Unicode UTC meeting
  2. API slush - checkins by Friday(13) next week - get proposals to the design list by this friday (6th)
  3. Yoshito: proposal: shorten key/value in CLDR identifier, may impact ICU. ICU may need to support truncated form of locale IDs.
  4. language tags ↔ locale ids - error mode, what to do in invalid case
  1. shouldn't happen in normal case
  2. could have a 'validate' function - determine if everything is of valid length and structure.
  3. expect try/catch around every use of a language tag?
  4. Distinguish 'fail to parse' from 'parse, but unrecognized'.  ( invalid vs. unknown )
  1. Bugs!

2009/01/28

CLDR status- in progress, probably need to check on new features in 1.7 to make sure they get implemented on ICU. Will file umbrella "make sure everything is in sync" bug. (John)

4.1.2 (Steven, Yoshito)

  1. completed both C and J
  2. version number to 4.1.3 (J done / C - srl to do. ?)
  3. r/t transliterator test bumped,
  1. will disable those in C and J. Already a CLDR bug [srl to find #] to implement these in CLDR.
  1.  and dt pattern generator bumped -Yoshito sent out a note and will create ticket.
  2. code review! will finish review of 4.0 bugs, please work on later ones also.
  3. Next milestone - (4.1.3) M3 Feb 18 / API slush - large API changes/addition must be in by M3
  4. Fri Feb 13, API checkin slush deadline. 

4.4 plan (Yoshito)

  1. need to sync with CLDR 1.8
  1. CLDR could always have minor releases that update to newest Unicode (already do
  1. sync with Unicode 5.2? October 2009. Will want a milestone to include that version.  (6.0 slated for fall 2010)
  2. Annual major release? Apr-May timeframe,
    then 1-2 maintenance releases between?
  3. More frequent milestone?
  4. March 10, 2010 - proposed 4.4 GA / propose the date to CLDR project next week

User Guide: Go forward with move to Site version (also at http://userguide.icu-project.org/ -- needs links fixed), or withdraw?

  1. go forward!
  2. Mark: didn't work on icu site more, maybe in some weeks.

Bugs

2009/01/21

4.0.1 status - shipped!

4.1.2 progress

  1. J - fine
  2. C -
  1. exhaustive mode cintltst failures ( ucnvsel / propsvec ) (markus and michael working on)
  1. Update: other tests pass.
  1. noted: AIX/GCC fails to build in strict(bug filed)
  2. noted: VS 2003 optimized fails bidi get mirrored char (srl to note in readme)
  3. noted: srl will start weekly exhaustive build
  4. 4.1.2 was tagged.. when there's a fix for ucnvsel, will merge it in to the tag and retest.

timezone 2009a (Yoshito)

  1. 'handle exception' in tz2icu on Cygwin.. looking ?
  2. changes to zic? take a look at any differences since tzcode was put into icu.
  3. Yoshito and John to look at?

Yoshito: will change CodeMangler to default to Java 6.

  1. eclipse IDE needs to be changed from workplace IDE 5 to Java 6. Yoshito to send instructions.
  2. Otherwise, strange and spurious checkin differences!
  3. there is an ant target that will reset the source code to prepare for checkin.
  4. Yoshito will checkin project file update update

Bugs?

  1. Bugs!

Please look at your bugs to review.

EOM (end of meeting..)

2009/01/14

ICU 4.0.1 status (Steven, Yoshito)

  1. ICU4C revision: r25245
  2. ICU4J revision: r25219 / tagged
  3. 4.0.1C candidate in testing
  4. should go out tomorrow (Thu, Jan 15)
  5. Beta download page up today.

ICU 4.1.2 schedule - M2 upcoming next Wed (1/21)

  1. #6433 xiaomei - do not need to pull out the current work-in-progress code in trunk - if xiaomei has time, merge current updated code from branch to trunk
  2. Friday noon PST - once stop code submission to trunk
  3. Any check-in after Friday noon - send a note to icu-core list
  4. Let M2 out as scheduled
  5. Add M3 (Feb 18) to catch up missing features

ICU 4.2 Platforms

  1. C/J: IBM binary/testing plans (Steven, Yoshito)
  2. C: VC 2005 -> 2008?
  3. IBM test:

  1. Windows Vista (32-bit), Visual Studio 2005 2008 express

    Windows Server 2003 2008 (64-bit), Visual Studio 2005 2008AIX 6.1 VisualAge 9 (64-bit)
    Solaris 10 (64-bit), Sun Studio 12
    Red Hat Enterprise Linux 5 (32-bit), GCC 4.1.2
    '''+Red Hat Enterprise Linux 5 (64-bit), GCC 4.1.2
    HP 11iv3 IA64 (64-bit) - aCC A.6.14
  2. ICU4J
    Note: Class files are Java 1.3 compatible (javac -target 1.3)
  1. Windows Vista (32-bit) - IBM JDK 6 (32-bit)
  2. Solaris 10 (SPARC 64-bit) - Sun JDK 6 (32-bit)
  3. AIX 6.1 - IBM JDK 6 (64-bit)
  1. Drop MSVC6 - no support

Brief discussion about MSVC 6:

Upgrade Java source file syntax level

Reminder - closing milestone 4.0 / 4.1.1

  1. We still need several tickets reviewed for 4.0

Bugs

2009/01/07

  1. ICU 4.0.1 status (Steven/Yoshito)
  1. Steven and Yoshito to update
  1. C++ static cast (Andy)
  1. Andy to try some in a branch - IBM team tries it with various compilers
  2. If it works, we will use them in new code.. Probably not trying to update everything at one time. 
  1. ICU Numbering System proposal (John)
  1. John to send out the proposal shortly.
  2. Force latin supported?
  1. The proposal is more comprehensive
  2. Yes? - Can be implemented like this way.
  3. How can we handle alternate decimal symbols (decimal point, grouping separator), not only for digits (0-9)?
  1. Bugs