Log Mac Meeting 12 September 2007

From Apache OpenOffice Wiki
Revision as of 14:35, 12 September 2007 by Ericb (Talk | contribs)

Jump to: navigation, search

Return to Previous meetings page : Previous_Mac_Meeting_logs

Return to September Meetings

[15:02] PhilippL meeting time ?

[15:02] shaunmcdonald PhilippL: +1

[15:03] PhilippL do we have an agenda ? the wiki entry is outdated ...

[15:03] * ericb2 didn't foudn the time to add entries. Sorry

[15:04] PhilippL traditionally we start with 1. Welcome new devs joining Mac OS X port

[15:04] PhilippL So any new people on board ?

[15:05] PhilippL ... seems not so.

[15:05] paveljanik Hi, I hope to be back on board after 2.3 8)

[15:06] PhilippL next point ? If there is one ?

[15:06] * mav_eric is now known as mav_eric_away

[15:07] paveljanik next point is the next meeting ;-)

[15:07] paveljanik what is the status of aqua% cwses?

[15:07] PhilippL aquavclcarbonfixes is RfQ

[15:07] PhilippL aquavcl03 is progressing. Currently I'm working on cocoafying printing.

[15:08] PhilippL However Cocoa's print model is patently unsuited to our own (contrary to the carbon one).

[15:08] paveljanik ups

[15:08] PhilippL s/unsuited/mismatched/

[15:09] ericb2 PhilippL: I still have the discussion we had. Did you continue the same way ? Create a new interface in offapi ..etc ?

[15:10] PhilippL I'm not talking about dialogs, I'm talking AquaSalInfoPrinter and AquaSalPrinter, sorry.

[15:10] ericb2 PhilippL: ok, I was confused

[15:10] PhilippL The native dialog things are not on my personal agenda right now.

[15:11] PhilippL ericb2: I didn't explain myself enough, sorry.

[15:11] ericb2 PhilippL: no problem, really :)

[15:13] PhilippL After printing only the application events (printing, open files) are missing in cocoafication of vcl, we should be as far as the carbon port was then.

[15:13] PhilippL modulo bugs of the new implementation of course ;-)

[15:13] PhilippL also outside of vcl nothing is cocoafied yet AFAIK.

[15:13] paveljanik hdu_hh: #i81231# is in cws with target 2.4 but has target 2.3.1?

[15:13] ericb2 PhilippL: are all events implemented ?

[15:14] IZBot framework DEFECT VERIFIED FIXED P3 System file-picker can't be used http://qa.openoffice.org/issues/show_bug.cgi?id=81231


[15:14] PhilippL ericb2: depends on what events you mean ?


[15:14] hdu_hh paveljanik: wrong target ;-)

[15:14] PhilippL ericb2: the application events are not yet, all view and window events are.

[15:15] paveljanik hdu_hh: ok :-)

[15:15] ericb2 PhilippL: ok. Do you have more infos about the getFrame() issue ?

[15:15] PhilippL ericb2: getFrame ?

[15:17] ericb2 PhilippL: in Cocoa version, add anything in equation editor causes a crash -> getFrame() is always on top of the stack

[15:17] PhilippL then we should fix that.

[15:17] ericb2 PhilippL: huu .. sorry, I believed the bug was already filed in some issue

[15:18] ericb2 PhilippL: shall I create one ?

[15:19] hdu_hh ericb2, PhilippL : I'm sure there is one... let me check

[15:19] PhilippL i don't see one in the current list, so please do.

[15:19] hdu_hh PhilippL: i80695

[15:19] ericb2 still for Cocoa version, I have proposed some little fixes about leaks in saltimer and salmenus

[15:21] PhilippL ericb2: any patches ?

[15:21] paveljanik PhilippL

[15:21] ericb2 PhilippL: I committed them already

[15:21] PhilippL ericb2: I think we talked about the menu stuff already, yes ?

[15:21] ericb2 PhilippL: I have another one for sal, but it can wait

[15:22] ericb2 PhilippL: yes we did

[15:22] PhilippL if it's already committed, then that's fine.

[15:22] ericb2 PhilippL: I still have a lot of crashes in menu too, with or without my patch. Maybe there is another cause

[15:22] ericb2 PhilippL: but sure I fixed some leaks

[15:23] PhilippL Funny, I don't get any. But then I never use the formula editor.

[15:24] ericb2 PhilippL: I started my courses using aquavclcarbonfixes, and no problem so far :)

[15:25] * ericb2 glad to have a working office suite ^^

[15:25] PhilippL So is there a next point ?

[15:25] ericb2 I'd like to discuss cws integrations

[15:25] PhilippL yes ?

[15:26] ericb2 macaddressbook01, macosxspellchecker, aquavclcarbonfixes ..

[15:26] ericb2 is it ok ? not ok ?

[15:26] PhilippL I think aquavclcarbonfixes is on its way. I'd like to resync aquavcl03 as soon as it is in.

[15:27] ericb2 Florian Heckl wrote me he has a new spotlight plugin

[15:27] ericb2 written from scratch

[15:27] ericb2 the plugin is under test since some times, and there is no known issue so far

[15:27] PhilippL cool

[15:27] ericb2 would be great to add it

[15:27] paveljanik macaddressbook01 is approved by QA, so it only has to be nominated.

[15:28] ericb2 paveljanik: great

[15:28] ericb2 no news about Drag and drop ?

[15:28] paveljanik spellchecker is still in RfQA, so no need to discuss about integration now.

[15:28] hdu_hh ericb2: sorry, no news about D&D

[15:28] hdu_hh :-(

[15:28] ericb2 hdu_hh: maybe Tino needs help ?

[15:29] ericb2 hdu_hh: or at least be two people to work on it could help ?

[15:30] PhilippL spellchecker has a rather red tinderbox state on Mac/X11

[15:31] ericb2 PhilippL: it is not made for X11, but maybe something is not correctly protected ?

[15:31] hdu_hh ericb2: Certainly. But AFAIK getting to know the D&D implementation details for outsider probably takes longer than having it implemented by the expert

[15:32] ericb2 hdu_hh: well, describe what has to be done should help too

[15:32] ericb2 hdu_hh: here we are stalled, and we have no info

[15:34] * dave_largo (n=drichard@64-132-254-110.static.twtelecom.net) has joined #ooo_macport

[15:34] ericb2 PhilippL: do you have an URL for the red logs with X11 build ?

[15:34] paveljanik ericb2: spellchecker is written without taking care about other platforms...

[15:34] PhilippL ericb2: http://go-oo.org/tinderbox/gunzip.cgi?tree=macosxspellchecker&brief-log=1189282268.7201

[15:34] paveljanik after looking at lingucomponent/source/spellcheck/macosxspell/makefile.mk:

[15:34] paveljanik +.IF "$(GUI)"=="WNT"

[15:34] paveljanik +ULINGULIB=libulingu.lib

[15:34] paveljanik +.ENDIF # wnt

[15:35] paveljanik It is nonsense to set it RfQA IMO

[15:35] ericb2 paveljanik: when did you see that ? today ?

[15:35] paveljanik ericb2: just now.

[15:36] ericb2 PhilippL: thank you

[15:38] ericb2 ok, who is volunteer to help mikesic to fix the portability issues in macosxspellchecker ?

[15:38] PhilippL tl should see that ?

[15:39] ericb2 PhilippL: tl is not there if I remember correctly

[15:39] PhilippL so ?

[15:39] PhilippL He had concerns about the separation from the hunspell checker anyway, are those solved ?

[15:39] ericb2 PhilippL: I don't know. Waiting, we can prepare patches, no ?

[15:45] PhilippL ping ?

[15:46] PhilippL anybody still out there ?

[15:46] ericb2 PhilippL: me :)

[15:46] paveljanik yup

[15:47] PhilippL so next point ?

[15:47] hdu_hh OOoCon? See you guys next wednesday!

[15:48] ericb2 hdu_hh: yes :)

[15:48] shaunmcdonald hdu_hh: less than a week to leave

[15:48] ericb2 splashscreen_OOoCon7.png

[15:48] ericb2 8)

[15:48] * ericb2 still has to finish the second part

[15:48] ericb2 FYI : http://eric.bachard.free.fr/mac/OOoCon2007/OOoCon2007_topical_item_ericb.odp

[15:48] hdu_hh ericb2: beautiful!

[15:49] ericb2 hdu_hh: don't exagerate :)

[15:50] * mikesic (n=mikesic@adsl-70-242-31-206.dsl.hstntx.swbell.net) has joined #ooo_macport

[15:50] * ChanServ gives channel operator status to mikesic

[15:50] PhilippL mikesic: we were talking about CWS macosxspellchecker

[15:51] mikesic PhilippL: Hi

[15:51] PhilippL mikesic: hi

[15:51] mikesic PhilippL: Can you summarize or send me a transcript?

[15:52] PhilippL mikesic: one thing is that the CWS does not seem to build on Mac/X11

[15:52] ericb2 paveljanik: do you think this is more correct: http://eric.bachard.free.fr/mac/aquavcl/patches/aquavcl03_september07/ spellchecker/macspell.diff

[15:52] PhilippL mikesic: I think there may be some adjustments to be made to the makefiles.

[15:52] mikesic PhilippL: I am sure you are correct.

[15:53] PhilippL mikesic: also I think tl had some concerns over service naming and separating the macosx spellchecker from hunspell. I don't know the details, could you tell us ?

[15:53] paveljanik ericb2: First, there should be big .IF GUIBASE == aqua or at least OS==MACOSX...

[15:54] paveljanik then you can do whatever you want to inside the .IF case, but please print "Nothing to be done for OS $(OS)..." in .ELSE

[15:55] ericb2 paveljanik: are you ok we build only for aqua ?

[15:55] mikesic PhilippL: Originally the mac spellchecker was in the same lib with the "SpellChecker" class, but they were separated. That could be it, but otherwise I don't know.

[15:55] paveljanik ericb2: if we can only test for aqua, sure ;-)

[15:55] paveljanik but it would be better to have it in all Mac versions...

[15:55] ericb2 paveljanik: I'm not sure it will work on X11

[15:56] paveljanik ericb2: depends on how it is done...

[15:56] paveljanik mikesic: ?

[15:56] mikesic paveljanik: yes?

[15:56] paveljanik mikesic: does your spellchecker code work in X11 as well?

[15:56] paveljanik does it use any native GUI?

[15:57] mikesic paveljanik: If we can link to the cocoa framework, there is no technical reason why not.

[15:57] mikesic paveljanik: I can try the X11 build and see what needs to be changed.

[15:58] mikesic paveljanik: No, It uses the OOo gui exclusively.

[15:59] PhilippL Then maybe it's just the makefile that needs to be changed.

[16:00] ericb2 PhilippL: I think so

[16:00] PhilippL from the tinderbox log it seems that the files are not compiled with objective C support.

[16:01] paveljanik I was not able to get the build logs from it 8)

[16:02] ericb2 PhilippL: accordingly to solenv/inc/unxmacx.mk, X11 version does not link with Cocoa framework

[16:03] PhilippL yes, one would need to add that in lingucomponent/source/spellcheck/macosxspell/makefile.mk for X11 case,

[16:03] * jsi_sun sorry for being late had to fix some environment issues with Mac OS X ;-)

[16:04] ericb2 PhilippL: can we separate Aqua macosxspellchecker implementation and add X11 version later ?

[16:04] PhilippL sure, just change the makefile so only on aqua any compilation is done.

[16:05] jsi_sun ericb2: then the CWS will get back to "New"?

[16:05] ericb2 jsi_sun: yes, unfortunaly :/

[16:05] jsi_sun All: Do we have an agreement of doing so?

[16:05] ericb2 jsi_sun: there is a portability issue

[16:05] PhilippL actually since no code is changing on aqua, why should we ?

[16:05] ericb2 paveljanik: what do you think ?

[16:06] ericb2 PhilippL: I'll follow you

[16:06] paveljanik ericb2: it definitely breaks other platforms now, so it definitely has to be brought back to New.

[16:06] ericb2 paveljanik: ok, then do you see other issues ?

[16:06] paveljanik first please make it build-clean than QA on it can start.

[16:06] PhilippL paveljanik: why ? there's nothing new to test, just a change to the makfile to NOT compile on other platforms ?

[16:06] jsi_sun paveljanik: +1

[16:07] mikesic paveljanik: ericb2 : In the makefile the line .IF "$(GUIBASE)"=="aqua" needs to change to allow X11 and aqua, can you tell me what it should be?

[16:07] * cloph_away is now known as cloph

[16:07] cloph Talking about macosxspellchecker?

[16:07] mav_eric_away mikesic mmeeks

[16:07] paveljanik PhilippL: sorry, but I know how programmers take care about other plaftoms ;-)

[16:07] PhilippL yes.

[16:07] ericb2 mikesic: .IF "$(OS)"=="MACOSX"

[16:07] cloph If so, ignore the Fedora x86_64 status - the cws is based on a pre-unomacli64 MWS...

[16:07] ericb2 mikesic: means both Aqua and X11. Is that what you need ?

[16:08] mikesic ericb2: Yes, thanks

[16:08] * cloph restricted the X86_64 tinderbox to milestones >=m222 in the meantime

[16:08] ericb2 mikesic: are you sure it will build and work on X11 ? This will force to do a twice QA

[16:08] TrainedMonkey cloph: what I don't like in ignoring like this is that bridges are very early built and noone knows whether later there is no breakage. I would prefer people with red tinderboxes to resync

[16:08] paveljanik lets go for aqua for now and do X11 later...

[16:08] paveljanik much easier.

[16:09] mikesic paveljanik: ok

[16:09] cloph The breaker on Mac is due to a fink-update I did - fink did break some dependencies - I solved that by building without gtk since then, but for some reason something in that cws pulls in a dependency from /sw/lib (from fink)

[16:09] ericb2 @all: ok for aqua only ?

[16:10] paveljanik ... for now... yes.

[16:10] cloph TrainedMonkey: Yes, a resync would clearly give a correct view wrt to the X86_64 box, but right now, the status is a (maybe) false positive

[16:10] TrainedMonkey maybe indeed

[16:10] cloph At least the place where it breaks is not related to the cws (it matches a known issue that was fixed within that 64bit cws)

[16:11] cloph TrainedMonkey: What do you think of ignoring the "[javac] The serializable class DetailsAction does not declare a static final serialVersionUID field of type long" style "errors"?

[16:11] * ericb2 wondering: reset to New, then patch lingucomponent/source/spellcheck/macosxspell/makefile.mk and finally set again as RfQA. OK ?

[16:11] TrainedMonkey cloph: yeah, because it is a pain

[16:12] cloph (ignore = not flag them as errors, so that they won't be listed in the summarylog)

[16:12] paveljanik mikesic: I'll work with on on this, ok?

[16:12] TrainedMonkey but it might be some integer width problem

[16:12] mikesic paveljanik: sure, that's great. Thank you :)

[16:12] TrainedMonkey cloph: just make it ignored

[16:12] cloph From what I can tell, all those errors are in "external" modules (beanshell, hsqldb, jfreereport)

[16:13] paveljanik jsi_sun: have you already started your work on it?

[16:15] jsi_sun paveljanik: you mean carbonfixes.... yes

[16:15] jsi_sun paveljanik: macosxspellchecker - also yes

[16:16] * jsi_sun lag sorry

[16:16] paveljanik jsi_sun: spellchecker: ups. Will you accept my word that no code change was there in New/RfQA round? ;-)

[16:17] jsi_sun paveljanik: I will accept your words.

[16:17] * paveljanik doesn't want to stop jsi_sun for real QA work

[16:17] paveljanik jsi_sun: great.

[16:17] jsi_sun paveljanik: I am on the test case specification trip, have done the needed tests and doing tests on aquacarbonfixes

[16:18] paveljanik great.

[16:18] * jsi_sun INFO: first.bas passed 100%; topten.bas FAILED in macro security (aquacarbonfixes)

[16:19] * jsi_sun now the Sun internal QA database accepting 'MacOS' entries, so it is easier for me to validate things

[16:19] ericb2 jsi_sun: what failed ?

[16:19] ericb2 jsi_sun: I tested the first.bas manually ? Just because high level security ?

[16:19] jsi_sun ericb2: No that is reflected in the test case; it sets the security down to low

[16:20] jsi_sun ericb2: I have to analyze it

[16:20] ericb2 jsi_sun: ok

[16:20] ericb2 jsi_sun: here, default is no macro. I had to enable the security level to start the test

[16:21] ericb2 Next point ?

[16:22] jsi_sun ericb2: "framework/first/inc/smoketest.inc::Macro didn't finish in time (10 minutes)"

[16:23] ericb2 jsi_sun: no idea .. :)

[16:23] shaunmcdonald jsi_sun: is the Sun internal QA database only accepting aqua mac os entries?

[16:23] ericb2 jsi_sun: can this avoid the cws ?

[16:23] jsi_sun shaunmcdonald: no

[16:23] jsi_sun shaunmcdonald: but we won't test the X11 internaly

[16:24] jsi_sun ericb2: I will see

[16:24] ericb2 jsi_sun: to be more precise, I'd suggest aquavclcarbonfixes as the first QA for aqua , i.e. as reference

[16:24] jsi_sun ericb2: Have to look at an older build and have to decide if the CWS is the breaker or the MWS also has that problem.

[16:24] shaunmcdonald jsi_sun: so will the crash reporter be turned back on

[16:24] shaunmcdonald report submitter

[16:25] jsi_sun shaunmcdonald: I don't speak about the crash reporter. I am talking about the QA database (test case results)

[16:25] <shaunmcdonald> jsi_sun: ah, so that they are two completely separate things

[16:27] * mav_eric_away is now known as mav_eric

[16:28] <ericb2> Next point ?

[16:30] * ericb2 received a proposal from a Software Architect + great experience, proposing to contribute in Aqua QA. jsi_sun : interested ?

[16:30] * ericb2 will ask him to join #ooo_macport

[16:30] <mav_eric> every Mac QA is welcome

[16:30] <ericb2> mav_eric: sure :)

[16:31] <mav_eric> and qa.openoffice.org

[16:31] <ericb2> No other point ?

[16:31] <ericb2> I have a question: *who* can to tests under Leopard ?

[16:32] <mav_eric> with the WWDC build yes - I have no access to later builds

[16:32] * ericb2 ADC select ends the 20th of September ( I paid it myself) , and I don't want to pay again.

[16:33] <jsi_sun> ericb2: yes, sure interested

[16:33] <ericb2> jsi_sun: I'll ask him to join the channel and contact you and QA project

[16:33] <jsi_sun> shaunmcdonald: yes, complete different databases; we will talk about that on OOoCon because we will put the database into the public in the near future

[16:33] <shaunmcdonald> jsi_sun: 8-)

[16:33] <jsi_sun> ericb2: ok

[16:34] <ericb2> Nobody runs Leopard on the channel ?

[16:34] <paveljanik> there is no Leopard yet.

[16:34] <paveljanik> 8)

[16:34] <mav_eric> ericb2: I have only the very outdated WWDC build and no money to join ADC

[16:34] <ericb2> paveljanik: it is announced on Apple site ;)

Personal tools