[05:17:04] *** peda_ has joined #JNode.org
[05:40:47] <tgiphil> when are most people online here?
[11:08:04] *** peda_uni_ has joined #JNode.org
[13:15:49] *** loic has joined #JNode.org
[13:15:54] <loic> hi
[14:28:07] *** echelog-1 has joined #JNode.org
[14:39:27] *** Mazon has joined #JNode.org
[15:29:47] *** rommanio has joined #JNode.org
[15:34:47] <tgiphil> when are most people online here?
[15:35:13] <rommanio> I am.
[15:37:18] <rommanio> I think, other people are working.
[15:40:45] <tgiphil> do you work on jnode?
[15:41:45] <rommanio> I'm a novice :)
[15:41:46] <rommanio> do yoo have a question?
[15:43:50] <tgiphil> I work on the mosa project and thought I would introduce myself and possibility ask a few questions
[15:44:30] <tgiphil> no questions right now, I'm getting ready to head to work myself.
[15:44:49] <rommanio> ah? mosa project?
[15:46:06] <tgiphil> managed operating system alliance (for .NET)... very similar to jnode.
[15:47:01] <rommanio> oh! and Singularity... well :)
[15:49:02] <rommanio> May you link me the project?
[15:52:22] <tgiphil> www.mosa-project.org
[15:52:45] <tgiphil> well, we're not singularity - and bsd open source.
[15:52:50] <tgiphil> time for me to run...
[15:54:06] <rommanio> Yes,long days and nice nights for you.
[16:34:11] *** loic has joined #JNode.org
[16:42:47] <peda_uni_> for me it's the evening/night hours.. I'm online during the day, too, though I should work at that time :)
[16:50:37] <rommanio> It's time for me...
[16:50:46] <rommanio> good bye!
[16:52:03] *** rommanio has quit IRC
[16:52:52] *** loic has quit IRC
[16:53:10] *** loic has joined #JNode.org
[18:26:25] *** FabienD has joined #JNode.org
[19:25:39] *** loic has joined #JNode.org
[19:31:30] <FabienD> hi all
[19:37:55] *** ismael has joined #JNode.org
[19:38:06] <ismael> hi
[19:38:14] <FabienD> hi ismael
[19:40:20] <loic> hi
[19:44:29] <ismael> peda_:  ping?
[19:47:36] <ismael> peda_uni_:
[20:07:13] <FabienD> ismael: loic: do you know lyon JUG : http://www.lyonjug.org/ ?
[20:07:20] <ismael> FabienD: yes
[20:08:17] <loic> FabienD: yes
[20:08:57] <ismael> FabienD: Julien (co founder) is a former INSA student
[20:09:03] <FabienD> in one week, it's the second birthday of paris jug : they have booked a 500 sits room !
[20:09:33] <loic> benoit and me have assisted to some lyon jugs
[20:09:52] <FabienD> (usually, it's another place with "only" 250 sits)
[20:09:55] <loic> but new ones are to far from our school, so we don't anymore
[20:11:24] <FabienD> they needed more sits and moved :) ?
[20:13:28] <loic> no idea
[20:14:36] <loic> i will ask benoit to ask to one of the lyon jug former (they work in the same office)
[20:17:08] <loic> sorry for the bad translation
[20:17:32] <loic> i mean not a former but a jug leader
[20:22:45] <loic> is it possible to retrieve an object size?
[20:24:21] <FabienD> it's not in the header ?
[20:27:12] <FabienD> I guess there is the size in clear or there is pointer to the next object
[20:29:56] <loic> peda: any idea?
[20:30:07] <loic> peda_uni_: ping
[20:32:53] <ismael> loic: the size is stored in the object
[20:33:08] <ismael> in alloc there is:
[20:33:08] <ismael>             objectPtr.store(alignedSizeW, sizeOffset);
[20:34:23] <ismael> there should be a method which does its reading
[20:34:28] <ismael> else you can do it yourself
[20:34:38] <ismael> with this code:
[20:38:04] <ismael> ObjectReference.fromObject(YOUR_OBJECT).toAddress().add(
[20:38:04] <ismael>                 sizeOffset).loadInt();
[20:49:52] <loic> how do i know the sizeOffset?
[20:50:22] <ismael> it's in heapHelper
[20:53:56] <ismael> actually in VmDefaultHeap. Sorry
[21:00:35] <loic> but it depends of the slotSize
[21:01:50] <loic> ok
[21:01:59] <loic> i have found how to handle this
[21:03:14] <ismael> loic: there might be a method already implemented doing the job
[21:04:43] <ismael> peda_uni_: peda_ : ping?
[21:48:27] <peda_> re
[21:48:28] <peda_> hi
[21:53:15] <ismael> peda_: hi
[21:53:52] <ismael> peda_: i added the metadata to objects before the color flag (offsets -3 and -4)
[21:54:03] <ismael> but i have some crashs now
[21:54:37] <ismael> how to avoid to write to previous object when i write these values
[21:54:41] <peda_> but you're sure the object itself is large enough? i.e. the -3 and -4 slot _cant_ point into another object?
[21:54:54] <peda_> your alloc implementation has todo that
[21:55:09] <peda_> if you get an allocObject(int size=12) ...
[21:55:22] <peda_> you'd actually allocate 12+4*slotsize bytes
[21:55:36] <peda_> and return the pointer+4*slotsize back to java
[21:58:31] <ismael> i don't see it done for the 2 already existing flags
[22:11:55] <ismael> should i modify HEADER_SLOTS? replacing 2 by 4 in objectLayout.
[22:13:44] <peda_> no
[22:13:50] <peda_> one sec
[22:14:40] <ismael> ok
[22:24:54] <peda_> if you have a look at the original implementation
[22:25:08] <peda_> in VmDefaultHeap.java#alloc(VmClassType, int)
[22:25:16] <peda_> the int is the alligned object size
[22:25:29] <peda_> some lines below there is:
[22:25:31] <peda_> final Word headerSize = Word.fromIntZeroExtend(this.headerSize);
[22:25:58] <peda_> and if you have a look at setupHeap in the same class
[22:26:00] <peda_> you see:
[22:26:09] <peda_> final int headerSize = ObjectLayout
[22:26:09] <peda_>             .objectAlign((ObjectLayout.HEADER_SLOTS + 1) * slotSize);
[22:26:26] <peda_> so the actual size to allocate is 1 slot more than in HEADER_SLOTS
[22:26:43] <peda_> this extra int is used for the objects size in the current implementation
[22:31:04] <ismael> so you advise me to replace this 1 by 3?
[22:31:48] <ismael> although I defined two more flag slot in object layout?
[23:59:56] <peda_> yup

top