Vm-dev August 2009

vm-dev@lists.squeakfoundation.org

27 participants
87 discussions

Some VMMaker updates, CPP generation and Mantis 5238 fix
by David T. Lewis 14 Aug '09

14 Aug '09

I have made some changes in VMMaker to resolve Mantis 5238 (Integer bug on 64bit image/32bit VM, DateAndTime now ==> -4391-12-26T16:18:33+00:00), which is a long-standing issue that affects Squeak 64-bit images. VMMaker-dtl.133 adds support for C preprocessor statements in Slang. This has no effect on the generated code for the interpreter (until of course you use one of the new Slang methods), but the changes touch quite a few methods and are likely to be an annoyance for anyone who may be maintaining a separate VMMaker project. VMMaker-dtl.134 has a new implementation of #isIntegerValue: that resolves the Mantis 5238 concern. This does change generated the C code (cosmetically), so I added it in a separate Monticello change to allow sceptical readers to separately verify that the C preprocessor support is not causing problems. Dave

1 0

Re: renaming dprintf
by Bert Freudenberg 13 Aug '09

13 Aug '09

On 13.08.2009, at 23:45, Ian Piumarta wrote: > On Aug 13, 2009, at 2:16 PM, Bert Freudenberg wrote: > >> We could rename the Balloon3D DPRINTF to DPRINTF3D and the regular >> dprintf to DPRINTF. > > Yes, we could. Nonetheless... > > I thought dprintf, being a GNU extension, would have absolutely no > right to pollute the global namespace unless specifically requested > by #define _GNU_SOURCE before inclusion of any header file? Sounds > like a serious bug in glibc and/or the config of gcc on fedora 11 to > me. Renaming every application function as/when gnu decides they > own the name is not a scalable solution. I hope someone with a > fedora on which they can make a small test case is reporting this to > the relevant people. Actually, it seems to be POSIX: http://www.opengroup.org/onlinepubs/9699919799/functions/dprintf.html - Bert -

1 0

Re: [squeak-dev] Built 64 bit vm for 64 bit compaq ES40 alpha server running gentoo linux.
by David T. Lewis 12 Aug '09

12 Aug '09

Richard, Very nice. Thanks for the link to http://mypals.info/tikiwiki-2.1/tiki-index.php?page=SqueakForAlpha. I am CCing this to the vm-dev mailing list, where various interested parties (including Ian) are likely to be reading. Dave On Tue, Aug 04, 2009 at 08:39:01PM -0500, richard snow wrote: > http://mypals.info/tikiwiki-2.1/tiki-index.php?page=SqueakForAlpha > > Here is a page with the two vms I've made (source code). > > the 3.8 is a 64 bit that runs 64 bit images, the 3.10 is a 64 bit that > runs 32 bit images. > > > On Tue, Aug 4, 2009 at 8:08 PM, David T. Lewis<lewis(a)mail.msen.com> wrote: > > On Wed, Aug 05, 2009 at 10:40:38AM +1200, Michael van der Gulik wrote: > >> On Wed, Aug 5, 2009 at 10:30 AM, David T. Lewis <lewis(a)mail.msen.com> wrote: > >> > >> > On Tue, Aug 04, 2009 at 04:16:02PM -0500, richard snow wrote: > >> > > I was able to use vmmaker to generate a 64 bit vm for the Alpha. > >> > > However it can't run 32 bit images, so I'm kind > >> > > of stuck. Where could I get an image (3.10-6 hopefully). > >> > > >> > >> I'm guessing that this would be a 64-bit image: > >> > >> http://squeakvm.org/squeak64/dist3/ > >> Squeak64-3.8g-6548.image.tar.gz<http://squeakvm.org/squeak64/dist3/Squeak64-3.8g-6548.image.tar.gz> > >> > >> But it's (IIRC) unsupported and probably unstable. 32-bit images are better > >> supported and the garbage collector isn't really designed for images bigger > >> than 4GB. > > > > There's nothing unstable about that image, but unfortunately it does not run > > on today's VMs (http://bugs.squeak.org/view.php?id=5239). Personally I use a > > descendent of that original 64-bit image (mainly for testing VMs) but as far > > as I know nobody is actively working on this. > > > > In any case, it's of acedemic interest in this case, because what Richard needs > > is a 64-bit VM to run normal 32-bit images. > > > > Dave > >

2 1

Invitation to submit to VMIL'09 [1 week extension]
by Michael Haupt 11 Aug '09

11 Aug '09

Dear Colleagues: It gives us great pleasure to invite you to submit your contributions to **the third international workshop on Virtual Machines and Intermediate Languages (VMIL 2009)**, which will be co-located with OOPSLA 2009. This workshop is a forum for research in virtual machines (VM) and intermediate languages (IL). It is dedicated to identifying programming mechanisms and constructs that are currently realized as code transformations or implemented in libraries but should rather be supported at VM and IL level. -------------------------------------------------------------------- **Invited Talks** VMIL 2009 will feature talks by Steve Blackburn, Vivek Sarkar and Doug Simon (tentative). -------------------------------------------------------------------- **Due Dates** Full papers are due Aug 22, 2009. Abstract submission is recommended but not required. -------------------------------------------------------------------- **PC** We have once again assembled an excellent program committee which consists of Eric Bodden, Alex Buckley, Andreas Gal, Doug Lea, Stefan Marr, Filip Pizlo, Andreas Sewe, Jan Vitek, and the organizers. -------------------------------------------------------------------- For more details please see: http://www.cs.iastate.edu/~design/vmil/ We look forward to your submissions to VMIL 2009. Best regards, Hridesh Rajan, Christoph Bockisch, Michael Haupt, and Robert Dyer VMIL 2009 Organizers

1 0

How to set up VMMaker class>>updateFromServer ?
by David T. Lewis 11 Aug '09

11 Aug '09

Can someone point me to a MonticelloConfigurations for complete dummies page? I'm interested in setting up something in VMMaker so we can do "VMMaker updateFromServer" with the result that VMMaker and some set of externally maintained plugins will all be updated to their latest versions. I like the way the Squeak trunk does this, so it would be good to copy that approach if possible. I am assuming that I want something like this: "MCMcmUpdater updateFromRepositories: #( 'http://squeaksource.com/VMMaker' 'http://squeaksource.com/OSProcessPlugin' 'http://squeaksource.com/AioPlugin' 'http://squeaksource.com/XDCP' )" But I don't know how to set up the actual configurations (and I don't want to mess up the repositories trying to figure it out). I tried reading the class comments in MonticelloConfigurations, but that didn't help much ;) Any pointers appreciated. TIA, Dave

2 2

[Bug] Bug in BitBlt (rgbMul sets alpha to 0)
by Juan Vuletich 11 Aug '09

11 Aug '09

Hi Folks, I found that rule 37 sets alpha to zero if the destination is 32 bit. This affects my antialiased StrikeFonts. The following f1 := Form extent: 8@8 depth: 32. f1 fillColor: (Color r: 1.0 g: 1.0 b: 0.7 alpha: 0.7). f2 := Form extent: 8@8 depth: 32. f2 fillColor: (Color r: 0.5 g: 1.0 b: 0.7 alpha: 0.9). bb :=(BitBlt toForm: f1) sourceForm: f2; combinationRule: 37; copyBits. (f1 colorAt: 2@2) answers Color transparent when it should answer (TranslucentColor r: 0.501 g: 1.0 b: 0.494 alpha: 0.627). This is related to the stuff I fixed in April, but it is a different bug in a different method, that I didn't find that time: #partitionedMul:with:nBits:nPartitions: ignores the last argument, and acts as if it was always 3. The fix I propose is as follows: partitionedMul: word1 with: word2 nBits: nBits nPartitions: nParts "Multiply word1 with word2 as nParts partitions of nBits each. This is useful for packed pixels, or packed colors. Bug in loop version when non-white background" | sMask product result dMask | sMask := maskTable at: nBits. "partition mask starts at the right" dMask := sMask << nBits. result := (((word1 bitAnd: sMask)+1) * ((word2 bitAnd: sMask)+1) - 1 bitAnd: dMask) >> nBits. "optimized first step" nParts = 1 ifTrue: [ ^result ]. product := (((word1>>nBits bitAnd: sMask)+1) * ((word2>>nBits bitAnd: sMask)+1) - 1 bitAnd: dMask). result := result bitOr: (product bitAnd: dMask). nParts = 2 ifTrue: [ ^result ]. product := (((word1>>(2*nBits) bitAnd: sMask)+1) * ((word2>>(2*nBits) bitAnd: sMask)+1) - 1 bitAnd: dMask). result := result bitOr: (product bitAnd: dMask) << nBits. nParts = 3 ifTrue: [ ^result ]. product := (((word1>>(3*nBits) bitAnd: sMask)+1) * ((word2>>(3*nBits) bitAnd: sMask)+1) - 1 bitAnd: dMask). result := result bitOr: (product bitAnd: dMask) << (2*nBits). ^ result If you agree, I'll add an issue to Mantis with the proposed fix. Cheers, Juan Vuletich

2 3

Squeak to Go
by K. K. Subramaniam 11 Aug '09

11 Aug '09

Hi Ian/Bert, etoys.sh has shaped up well as a smart launcher script that can launch squeak binary from any location (e.g. $HOME/.squeak, a mounted volume/share or USB flash memory chip). It can be merged into the unix/src tree. I propose that we generate the main binary as $bin/squeakvm and install this script (with some minor modifications) as $bin/squeak. This will help us get rid of /usr/local/lib prefix hard-wiring. Subbu

2 1

Unix VMs
by Bert Freudenberg 10 Aug '09

10 Aug '09

For Etoys-To-Go (http://wiki.squeakland.org/display/sq/Etoys+To+Go) I'd like to package a few more VMs. In particular the ones downloadable from http://squeakvm.org/unix/ - but I need to detect the system we're running on. Currently I'm using: echo `uname -s`-`uname -m` which on this box gives "Linux-i686". Could someone report what they get on BSD, Solaris, etc? - Bert -

5 9

Linux SecurityPlugin problem
by Bert Freudenberg 10 Aug '09

10 Aug '09

Segmentation fault 2032281096 ByteString>? 2031934628 ProjectLoading class>checkSecurity:preStream:projStream: 2030862272 ProjectLoading class>morphOrList:stream:fromDirectory:archive: 2030860572 ProjectLoading class>openName:stream:fromDirectory:withProjectView: [...] This looks very much like the problem on Windows http://squeakland.org/forums/viewtopic.php?t=4203 which I thought was traced to the stack balancing issue: http://bugs.squeak.org/view.php?id=7349 However, it supposedly was fixed in VMMaker-dtl.122 and Ian used vmm- dtl-126 for his latest release (I'm testing the pre-compiled 3.10-6 Linux binary). Any idea? - Bert -

1 1

Linux HostWindowPlugin does not work
by Bert Freudenberg 10 Aug '09

10 Aug '09

Hi Ian, the HostWindowPlugin in Squeak-3.10-6.i686-pc-linux-gnu does not work. It misses the generated primitives from src/plugins/HostWindowPlugin/ HostWindowPlugin.c which indeed is not in the unix tree. My working version is twice as large: [bert@fedora10 Linux-i686]$ ll HostWindowPlugin* -rw-r--r-- 1 bert bert 7376 2009-08-10 13:46 HostWindowPlugin -rw-r--r-- 1 bert bert 4048 2009-08-10 13:46 HostWindowPlugin.ian Could you fix that? - Bert -

1 0

← Newer
1
2
3
4
5
6
7
8
9
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

Vm-dev August 2009