Log Mac Meeting 4 May 2007

From Apache OpenOffice Wiki
Jump to: navigation, search

Return to Previous meetings page : Previous_Mac_Meeting_logs

Return to April Meetings Mac_meetings_May_2007

[22:00] <ericb2> Can we start ?

[22:00] <shaunmcdonald> ericb2: yes

[22:00] <ericb2> 1. Welcome new devs and devs from Sun joining Mac OS X port

[22:00] <ericb2> :)

[22:00] <shaunmcdonald> the agenda is at http://wiki.services.openoffice.org/wiki/MacOSXPortMeetings

[22:01] <ericb2> shaunmcdonald: thanks

[22:01] * cloph_away is now known as cloph

[22:01] <ericb2> as usual , the first point is dedicated to welcome the new devs

[22:01] <ericb2> today is not usual, becasue we have to welcome Sun devs

[22:02] <paveljanik> we want to :-)

[22:02] <PhilippL> Oh, I'm here :-)

[22:02] paveljanik PhilippL plipli

[22:02] <PhilippL> Hello everybody.

[22:03] <paveljanik> Looks like _Nesshof_ finally trashed Windows as well? ;-)

[22:03] <ericb2> PhilippL: can you please shortly present you ?

[22:03] paveljanik PhilippL plipli

[22:03] * cloph wonders whether anybody commented to my nas question/rant since I have been disconnected...

[22:03] <_Nesshof_> paveljanik: not really, I'm using all the platforms where OOo runs

[22:04] * PhilippL works for Sun/StarOffics/StarDivision for nearly ten years. basically always on the platform GUI stuff, and here mostly for the Unix platforms.

[22:04] * _Nesshof_ admits that he runs now Vista

[22:05] * Cremlae (n=Cremlae@dialup-85-141.colorado.edu) has joined #OOo_macport

[22:05] <PhilippL> Apart from vcl I work on printing clipboard, PDF export, scanning and have left fingerprints at a dozen other places.

[22:05] <plipli> hi Cremlae

[22:05] <paveljanik> PhilippL: glad to have you on board!

[22:05] <Cremlae> plipli: Hi!

[22:06] <PhilippL> I have basically no mac experience, but I think I can help with integration into general vcl.

[22:06] <plipli> PhilippL: great to have a vcl expert with us

[22:06] <PhilippL> HDU on the other hand is THE font expert inside OOo I would say.

[22:06] <ericb2> PhilippL: sure you will : your vcl knowledge is IMHO impressive

[22:06] <plipli> we need vcl knowledge, we ll bring mac osx experience

[22:07] <PhilippL> The aqua ATSUI code is already contributed (at least partly ?) by him and he will certainly continue to work on that.

[22:07] <plipli> hdu has helped eric and me for atsui

[22:07] <PhilippL> I guess that's about it about us ?

[22:07] <ericb2> FYI , Herbert advertised me he cannot attend the meeting

[22:07] paveljanik PhilippL plipli

[22:07] paveljanik PhilippL plipli

[22:08] * fheckl (n=florianh@p57B55D99.dip.t-dialin.net) has joined #ooo_macport

[22:08] <plipli> With hdu tips, I was able to inplement beginning parts of salatslayout

[22:08] <ericb2> PhilippL: thank you very much

[22:08] <shaunmcdonald> fheckl: hello, PhilippL has just introduced himself

[22:09] <plipli> hi fheckl

[22:09] <fheckl> ok, I read his intro mail, hi everyone

[22:10] <paveljanik> other new devs?

[22:10] <ericb2> paveljanik: I think so

[22:12] <impad> yes, I am a standalone programer with some experience with OSX dev

[22:12] <impad> but a newcomer on the project :)

[22:12] <ericb2> impad: great ! be welcome :)

[22:13] * Fridrich (n=fridrich@26-34.1-85.cust.bluewin.ch) has joined #ooo_macport

[22:13] * ChanServ gives channel operator status to Fridrich

[22:13] <impad> ericb2 : thank you

[22:13] <ericb2> impad: do you have Carbon API experience ?

[22:14] * damiend (n=damiend@fau42-2-82-239-192-212.fbx.proxad.net) has joined #ooo_macport

[22:14] <ismael_> hi damiend

[22:14] <damiend> ismael_: ih

[22:14] <impad> ericb2 : unfortunately, I'm much more experienced in Cocoa

[22:14] <damiend> ismael_: i'm reading the log

[22:15] <ericb2> impad: if you have Mac experience, you can help us

[22:15] impad ismael_ IZBot

[22:15] <PhilippL> Question: Is the Cocoa implementation based on Carbon ?

[22:15] <impad> ericb2 : at least, I'll try ;)

[22:16] <fheckl> PhilippL; given Cocoa's past, I doubt it

[22:16] <ericb2> PhilippL: no, this is another API ( objective C, simple inheritance, dynamic typing ..etc) . Most of the time Carbon "compatible" , but not always

[22:16] <paveljanik> PhilippL: some APIs are, some are not.

[22:17] <PhilippL> So is there any real danger of Apple end-of-lifing Carbon ?

[22:17] * doctype (n=martin@p57a23df8.dip0.t-ipconnect.de) has joined #ooo_macport

[22:18] <fheckl> PhilippL: not in the next few years

[22:18] <plipli> ericb2: i don t agree

[22:18] <tinor_> PhilippL: In principal yes, but I think its almost the same as with win32 on Windows, they can really afford to end-of-lifing it

[22:18] <tinor_> PhilippL: Without loosing lots of old apps

[22:18] <thorsten> which other apps are based on Carbon?

[22:19] <PhilippL> Ok, i just asked, because that was of the concerns in the Blog-Comments

[22:19] <thorsten> mozilla, I guess?

[22:19] <tinor_> #can#can't

[22:19] <damiend> thorsten: old CS2 Adobe and Office 2004

[22:19] <damiend> but often into PowerPPC arch

[22:19] <paveljanik> next point? ;-)

[22:19] <plipli> cocoa and carbon are mixed in macosx apis

[22:19] <tinor_> thorsten: AFAIK many of the Adobe apps

[22:20] <ericb2> Other devs ?

[22:20] <plipli> we have to focus on

[22:20] <plipli> paveljanik: no

[22:20] <plipli> a sec

[22:20] <tinor_> thorsten: Even MS Office is based on Carbon, though I've heard they are trying to switch to Cocoe

[22:20] <plipli> cocoa can use carbon and carbon can call cocoa

[22:20] <plipli> we don t need to switch truly

[22:21] <paveljanik> good example will be spell checker integration for us ;-)

[22:21] <plipli> apple release api for both cocoa and carbon

[22:21] <plipli> becaus cocoa is too high level

[22:21] <plipli> for specific dev every mac coder use part of carbon apis

[22:21] <ericb2> plipli: and this is expensive .. maintain only one would simplify apple life

[22:22] <plipli> all this will be much clearer zith core grpahics, core text

[22:22] <ericb2> Next point ?

[22:22] <shaunmcdonald> one quick thing

[22:22] <plipli> ericb2: the transition has alreqdy begun

[22:22] <shaunmcdonald> the news of the sun devs is reaching far

[22:22] <shaunmcdonald> there have been around 2 dozen blog posts or more so far

[22:22] <plipli> we have to focus on being Core * co;patible

[22:23] <shaunmcdonald> this is a *lot* compared to an ooo release or any other mac related ooo news

[22:23] <obr> paveljanik: is somebody actively working on spellchecker integration ?

[22:23] <paveljanik> obr: no.

[22:23] <obr> paveljanik: o.k., thanks.

[22:23] <damiend> plipli: what do you mean by "Core compatible ? "

[22:24] <shaunmcdonald> this have been received mostly positively :-) though there have been a few comments on "Just use NeoOffice"

[22:24] <plipli> if we continue to use recent apple code, sample and core grpahics, core text etc

[22:24] <plipli> we ll be safe

[22:24] <ericb2> plipli: +1

[22:24] <ericb2> plipli: this is my opinion too

[22:24] <plipli> I ll get info more precise from apple engineers

[22:24] <plipli> that s one of the main reason to go to wwdc

[22:25] <plipli> and fetch info at the source

[22:25] <fheckl> plipli: yes, just think of QuickDraw deprecation in favor of CoreGraphics/Quartz

[22:26] * Fridrich (n=fridrich@26-34.1-85.cust.bluewin.ch) has left #ooo_macport

[22:26] <ericb2> Next point is :

[22:26] <plipli> fheckl: cocoa is only a framework historically from nextstep

[22:26] <ericb2> 2. A point on Aqua port

[22:26] <plipli> fheckl: now Tiger and leopard are Core* based

[22:27] <plipli> damiend: we have to continue following apple s guide to Core oriented progrmming : using API of Core frameworks

[22:27] <plipli> ericb2: +1

[22:28] <ericb2> The Aqua port now

[22:28] <ericb2> our first objective is to provide a real Mac application

[22:29] <ericb2> since some times ,we did big progress, enough to define a timeline and a Top 10 of issues before to provide an "alpha"

[22:30] <ericb2> The timeline : http://porting.openoffice.org/mac/timeline.html

[22:30] <ericb2> as you can see, we are a bit in late

[22:30] <ericb2> Top 10 of issues: http://porting.openoffice.org/mac/news/2007/20070203toptenbeforealpha.html

[22:30] <ericb2> 5 are ( please correct me) fixed

[22:31] <ericb2> 3 over 5 have a workaround ( we know the " what", not the exact "why")

[22:32] <ericb2> During 2nd Hamburg porters meeting (dec2006)

[22:32] <shaunmcdonald> ericb2: very good point about the timeline

[22:32] <ericb2> Stephan Shaefer wrote a document describing the most important points to solve

[22:32] <ericb2> http://eric.bachard.free.fr/mac/MacPortersMeeting_Ham_07/may04/aquaport-roadmap.pdf

[22:33] <ericb2> I just updated the document ( the old one is : http://eric.bachard.free.fr/mac/MacPortersMeeting_Ham_07/ )

[22:34] <ericb2> I proose to discuss the content, point by point, and define objectives

[22:34] <ericb2> s/proose/propose/

[22:34] <ericb2> sorry

[22:35] <PhilippL> The first point seems to be getPixel ?

[22:35] <ericb2> PhilippL: if I'm not wrong, yes

[22:35] paveljanik PhilippL plipli

[22:35] <ericb2> PhilippL: currently we return nSalColor =0

[22:36] <ericb2> PhilippL: and I wonder ..

[22:36] <PhilippL> The question is whether it is needed, the answer is: seldom, but in some places, yes. It should be possible to implement since you have getBitmap, so the hard way would be getBitmap and then get the pixel out of the bitmap.

[22:37] <paveljanik> or getBitmap with 1x1 rectangle...

[22:37] <ericb2> PhilippL: what is the main purpose of this method ?

[22:37] <ericb2> paveljanik: yes, this is what I had in mind

[22:37] paveljanik PhilippL plipli

[22:37] <plipli> PhilippL: for a general purpose, I think what we need and that you can teach us is : what is the goal of each mandatory function/method

[22:37] <ericb2> paveljanik: some CGRect

[22:37] <PhilippL> The most common example i found is getting the background color to blend with.

[22:37] <PhilippL> Arguably that is just bad implementation.

[22:38] <thorsten> yeah - Impress edit engine samples five points from the bg to calc its background color ;-P

[22:38] <PhilippL> One valid example: the eyedropper

[22:39] <PhilippL> There you pick a color out of the document to work with.

[22:39] <PhilippL> This cannot be implemented differently I think.

[22:39] <plipli> PhilippL: +1

[22:40] <thorsten> well - mid-term, all apps will have switched to aw's overlay buffer

[22:40] <thorsten> so, there's always a VDev with the currently visible content

[22:41] <PhilippL> Which will need getPixel :-) Also I'm not sure whether this will squat the eyedropper case.

[22:42] <thorsten> PhilippL: sure - but getPixel in a VDev should be fairly easy ;-)

[22:42] <PhilippL> Since with that you pick a color out of the document and save it e.g. for use in a color scheme.

[22:42] <PhilippL> thorsten: point taken.

[22:43] <PhilippL> However when the overlay will come to pass is unknown, so we better implement getPixel using getBitmap just now.

[22:43] <PhilippL> If we don't need it anymore, it can go.

[22:43] <thorsten> yep

[22:45] <ericb2> maybe this is enough infos ?

[22:45] <PhilippL> please tell me if I'm too noisy :-)

[22:45] * Nesshof_ (n=chatzill@p5B05C866.dip.t-dialin.net) has joined #ooo_macport

[22:45] <ericb2> PhilippL: no, at all

[22:46] <ericb2> PhilippL: we are glad to learn more about vcl

[22:46] * mreimer_ can tell it's going to be very helpful having sun devs helping

[22:47] <PhilippL> Next point on the list is Beziers ? That's optional. I would suggest leaving this empty for now.

[22:47] <ericb2> ok, agreed

[22:47] <ericb2> next is drawBitmap

[22:49] <ericb2> in AquaSalGraphics, they are 3 drawBitmap() , and one is not implemented

[22:50] * _Nesshof_ has quit (Remote closed the connection)

[22:51] <PhilippL> That's the one with transparent color ?

[22:51] <paveljanik> only the one with transparent color is not implemented.

[22:51] paveljanik PhilippL plipli

[22:51] <paveljanik> 8)

[22:51] <paveljanik> yes

[22:51] <thorsten> ah, that mask color one. can surely be emulated by the mask bmp one

[22:51] <ericb2> PhilippL: exactly

[22:52] <PhilippL> That one was used for toolbar icons, but nowadays these have alpha masks or bitmap masks since they are png files.

[22:52] <thorsten> PhilippL: we should even be able to make that method optional as well, and emulate in the indepent part

[22:52] <PhilippL> I cannot say for sure how often that one gets called.

[22:52] <PhilippL> thorsten: yes, since we can always create a one bit mask here.

[22:53] <ericb2> ok, noticed

[22:53] <ismael_> PhilippL: IMHO this functions is not called often, and maybe never called

[22:54] <ismael_> thorsten: +1

[22:54] <PhilippL> ismael_: I think you are right nowadays

[22:54] <ericb2> Next point does concern invert()

[22:54] * thorsten remembers some dark corners were that's still used

[22:55] <ericb2> we have two invert() one, works maybe not completely, the other one is not implemented

[22:56] * ybart (n=ybart@lns-bzn-27-82-248-23-125.adsl.proxad.net) has joined #ooo_macport

[22:56] <ericb2> all files are in salgdi.cxx FYI

[22:57] <ericb2> s/files/methods/

[22:57] <PhilippL> The difference for the second one is simply that it draws a Polygon instead of a rectangle. If the rectangle one works, the other should not be a big problem to implement I think.

[22:57] <ericb2> PhilippL: ok

[22:59] <ericb2> Next point is Font /text handling

[23:00] <PhilippL> I guess that is what HDU will be doing the next months :-)

[23:00] <PhilippL> Unless there is already somebody working on this ?

[23:00] <ericb2> i think so too :)

[23:00] <paveljanik> tinor_: plipli and tinor_ were working on this?

[23:01] tinor tinor_

[23:01] <tinor_> paveljanik: That's true

[23:01] <tinor_> I talked to HDU already and will hand over my changes to him

[23:01] <paveljanik> tinor_: great!

[23:01] <tinor_> I'm doing some small tests to see that my changes do not break the build

[23:02] <tinor_> I will then check them in

[23:02] <plipli> I ll talk with hdu

[23:02] <tinor_> So that HDU can pick them up and continue

[23:02] <paveljanik> I think we can (for now) simply skip this point :-)

[23:02] <plipli> I ve inplemented some part in salatslayout mainly for glyph handling and metrics for carets

[23:03] <plipli> paveljanik: +1

[23:03] <ericb2> Next point is native widget rendering

[23:03] <ericb2> slanativewidgets.cxx is concerned

[23:03] <plipli> we have to finish integration of hiframework

[23:04] <plipli> with hiframewok we ll be able to add native widgetd fastly

[23:05] <ericb2> this is the part Ismael will work for the Google Summer of code

[23:05] <plipli> YEP

[23:07] <paveljanik> SalGraphics now?

[23:07] <ericb2> SalGraphics

[23:07] * ybart has quit (heinlein.freenode.net irc.freenode.net)

[23:07] * doctype has quit (heinlein.freenode.net irc.freenode.net)

[23:07] * Cremlae has quit (heinlein.freenode.net irc.freenode.net)

[23:07] * sky has quit (heinlein.freenode.net irc.freenode.net)

[23:07] <PhilippL> Clipping ?

[23:08] <plipli> cipboard ?

[23:08] <ericb2> PhilippL: yes. the question is mine : I found several method twice defined

[23:08] <ericb2> clipping

[23:08] <plipli> PhilippL: what is clipping made for ?

[23:08] <PhilippL> Setting a clip region will confine drawing to that area

[23:09] <PhilippL> If you draw outside that region, nothing will happen.

[23:09] <PhilippL> E.g. during a paint event the drawing is clipped to the area that needs to be updated.

[23:11] <PhilippL> eric: Do you mean the clip methods on SalFrame and SalGraphics ? These have completely different meanings.

[23:11] <paveljanik> ericb2: twice?

[23:11] <paveljanik> PhilippL: same meanings on different objects, right?

[23:11] <PhilippL> The one in SalGraphics will clip painting operations.

[23:12] <PhilippL> The one in SalFrame will actually shape the window.

[23:12] <ericb2> PhilippL: ahh

[23:12] <ericb2> PhilippL: thanks

[23:12] <PhilippL> So you do not have a rectangular window but e.g. one with rounded edges.

[23:13] * doctype_ (n=martin@p57a23df8.dip0.t-ipconnect.de) has joined #ooo_macport

[23:13] <ericb2> PhilippL: so they are mandatory. that's what I missed

[23:14] <PhilippL> The ones on SalFrame do not get used much. However the clipping on SalObject, that does much the same is important for plugins and applets.

[23:14] <PhilippL> And of clipping on SalGraphics gets used heavily :-)

[23:15] <PhilippL> ericb2: Actually the only customer for shaped windows at the moment is the update notification window.

[23:15] <ericb2> PhilippL: not used afaik on Mac OS X

[23:16] <damiend> i've to go

[23:16] <damiend> 'd night all

[23:16] <PhilippL> It won't me missed for a while I think. However there are plans to move tool tip windows to a similar look, so over time we will want them.

[23:16] * damiend has quit ()

[23:17] <ericb2> the list is long

[23:17] <ericb2> Drawing modes now

[23:18] <PhilippL> I THINK these are mostly there because metafiles can contain them. I'm not aware of applications using that otherwise.

[23:18] <PhilippL> So the RasterOps can be contained in documents.

[23:19] <PhilippL> However they will not work everywhere anyhow, e.g. never on a printer.

[23:21] <ismael_> i have to go, bye

[23:22] <ericb2> ismael_: bye

[23:22] <paveljanik> Where is setting the XOR mode used in OOo?

[23:22] * ismael_ has quit ()

[23:23] <PhilippL> paveljanik: consider a document containg a metafile. that metafile graphics can contain XOR or one of the other ROPs. This will set the raster op for drawing.

[23:24] <paveljanik> so in short: for now, it is not critical?

[23:24] <PhilippL> yes, I think so. However later there may come some bugs in with documents that paint incorrectly.

[23:25] <paveljanik> ok

[23:26] <paveljanik> we can now skip to Event Loop I think.

[23:26] <ericb2> If the next ( Get Device REsolution BitCount ) is scond importance, I propose we jump to Window Handling

[23:26] <ericb2> paveljanik: or Event Loop

[23:26] <paveljanik> we have to update used API for Timers.

[23:27] <paveljanik> and we still do not properly refresh in the current cws, so probably some issues are still hidden here.

[23:27] <paveljanik> plipli: or were you able to see correct refresh already?

[23:29] <paveljanik> ok, Window handling now.

[23:29] * tinoarbeit (n=tinoarbe@e176091188.adsl.alicedsl.de) has joined #ooo_macport

[23:29] <ericb2> tinoarbeit: ?

[23:29] <ericb2> :)

[23:30] * sky (n=sky@roam.cybotech.net) has joined #ooo_macport

[23:30] * Cremlae (n=Cremlae@dialup-85-141.colorado.edu) has joined #OOo_macport

[23:30] <ericb2> Maybe we can continue the discussion later ?

[23:30] <tinoarbeit> ericb2: just got kicked off, rejoined..obviously with a different identity :)

[23:31] <paveljanik> ericb2: I think so.

[23:31] <plipli> +1

[23:31] <plipli> window and controls : hiframework based will rules ;-)

[23:31] <plipli> good night asll

[23:31] <plipli> all

[23:32] <Cremlae> plipli: Good night!

[23:32] <ericb2> ok, i propose to continue next week, and I'll put all these informations somewhere on the wiki

[23:32] <plipli> bye Cremlae

[23:33] <ericb2> PhilippL: no problem is I contiunue with a lot of questions at office time ?

[23:33] <ericb2> s/is I/if I/

[23:33] <PhilippL> ercib2: None at all :-)

[23:33] paveljanik PhilippL plipli

[23:33] <ericb2> PhilippL: thanks :)

[23:34] <PhilippL> s/ecib2/ericb2/ (it's getting late)

[23:34] <ericb2> PhilippL: for all

[23:34] <tinoarbeit> I think I will call it a day too now

[23:34] <obr> seems that PhilippL did not activate tab completion yet ;-)

[23:34] <ericb2> Quickly : WWDC weekly news

[23:34] <PhilippL> Have a nice weekend everybody.

[23:34] <ericb2> PhilippL: see you !

[23:35] <tinoarbeit> My aquavcl01 m211 build just finished without any problems

[23:35] <thorsten> same to you, PhilippL !

[23:35] * tinor_ has quit (Read error: 145 (Connection timed out))

[23:35] <ericb2> Only plipli and me will attend

[23:35] <plipli> yes

[23:35] * tinor_ (n=tinor@e176091188.adsl.alicedsl.de) has joined #ooo_macport

[23:35] <tinoarbeit> I think I can check in my font changes over the weekend

[23:35] <paveljanik> tinoarbeit: great!

[23:35] <plipli> bye

[23:35] <ericb2> and the money remaining will be reversed for bounties or something like that

[23:35] * plipli has quit ()

[23:35] <tinoarbeit> 'g night all

[23:35] * PhilippL has quit ("ChatZilla [Firefox]")

[23:35] <ericb2> I'll try to find the exact amount, but things are complicated

[23:35] * tinor_ has quit (Client Quit)

[23:35] * tinoarbeit has quit ("See you later")

[23:36] <thorsten> bye folks!

[23:36] <ericb2> eTickets are on

[23:36] <ericb2> thorsten: bye

[23:36] * thorsten is now known as thorsten_away

[23:37] <ericb2> Another point : Michael Sicotte contributes a lot , and does provide regular and quality code

[23:37] <ericb2> I propose to ask commit rights for him

[23:37] <ericb2> he really desserves it

[23:38] <paveljanik> +1

[23:38] <fheckl> ericb2: +1, but can't he ask for them himself?

[23:38] <fheckl> ericb2: :)

[23:38] <paveljanik> yes, he should do so in IZ ;-)

[23:38] <ericb2> fheckl: he will do. He just could not attend this evening

[23:40] <ericb2> Postponed : prepare OOoCon2007 , but take care : deadline is 1st of June, and it would be great to see a macximum of Mac porters attend Barcelona !

[23:40] <ericb2> and contribute or propose a conf or BoF .. etc

[23:41] <ericb2> Next meeting ?

[23:42] * ericb2 proposes Friday 11th of May same hour

[23:42] <ericb2> Questions  ?

[23:42] <paveljanik> next week, one hour earlier? ;-)

[23:42] <paveljanik> the first part of the meeting was productive, the second was not...

[23:43] <ericb2> paveljanik: we can try

[23:43] <paveljanik> probably better agreed on mailing list...

[23:43] <paveljanik> not all people are here.

[23:43] <ericb2> paveljanik: can you ask ?

[23:43] <paveljanik> in the ML? Sure.

[23:43] <paveljanik> (if none volunteers ;-)

[23:44] <ericb2> paveljanik: thanks

[23:44] <ericb2> End of meeting, see you !

Return to Previous meetings page : Previous_Mac_Meeting_logs

Return to April Meetings Mac_meetings_May_2007

Personal tools