- Pharo-users - lists.pharo.org

[UFFI] How to get a string from the a void pointer
by Dimitris Chloupis Nov. 7, 2016

Nov. 7, 2016

So I managed to port all C++ functions to pharo, which basically includes the following Not full definition of function open() lseek() close() mmap() I have tested them all, they all return proper values from inside Pharo and Pharo does not crash. it seems also that mmap works because it returns a void pointer as it should. This pointer points to the shared memory where I have put the string "hello" from another C++ program that has shared the memory . So how can from Pharo take the void pointer "(void*)@ 16rFFFFFFFF" and retrieve the string "hello" that it points to ? Essentially I want to dereference the pointer I tried to send the message fromCString to the void pointer (ExternalData) but it crashed Pharo. Another question, is it possible to cast a pointer ? (I do this with C++ programm) so can I cast from pharo the void pointer to a string pointer ? I have also no idea if the void pointer points to a proper memory address , the documentation says that all pointers live in head but I have no way to confirm this. I am so close to sharing memory between Pharo and C++ that I can smell it :D

1 1

Pharo desktop UI
by vikenti Nov. 7, 2016

Nov. 7, 2016

Hello, i'm newbie in Pharo, converted from Visual Works which is quite different from Pharo Smalltalk. I started to port my application (i've been working on for about three years) from VW (because of new personal user licensing limitations of Visual Works), but i have to solve some problems and need your advice. These ase some of the most important problems: 1. I need to rewrite the whole GUI from scratch. What UI framework to choose and where to get documentation? Please, don't send me to research code examples instead of docs - i do this during whole last week. The example should be complete and simple: one task - one example, all examples i've seen are either not applicable to real life or too complicated to understand. - I need simple window with menu (how to attach menu items to menu handlers???? - no answer in examples) - Toolbar with icons (how????????) - components (as i understand - Morphic subclasses, it is the simplest part) - when window opens i need handler to do some setup actions. Handler should fire after UI is built and my app's window is opened (how?????) - i need another handler, which fires when you try to close Pharo or application's window. Handler must check whether to close window or not and do allow or not to close my application. - how to programmatically close window?????? 2. The next task is to do some work on image load (i mean pharo image). I need some handler which fires immediately after image is loaded. How to include my handler in this chain???? My app saves data into the image so i need to manage this and auto-load _my_ application after startup. 3. How to drag-n-drop from the host operation system into pharo application and vice versa? I need to drag-n-drop a list of files from\to file browser. 4. Application deployment. How to strip the unnecessary code from the image? I've heard there is "small image" to add all-i-need-to-run. Where i can download it and how to use it???? -- View this message in context: http://forum.world.st/Pharo-desktop-UI-tp4921212.html Sent from the Pharo Smalltalk Users mailing list archive at Nabble.com.

6 16

Re: [Pharo-users] Speed up JPG images loading
by Sven Van Caekenberghe Nov. 7, 2016

Nov. 7, 2016

Hi Matteo, > On 5 Nov 2016, at 16:11, Matteo via Pharo-users <pharo-users(a)lists.pharo.org> wrote: > > > From: Matteo <matteob8(a)yahoo.it> > Subject: Speed up JPG images loading > Date: 5 November 2016 at 16:10:45 GMT+1 > To: "Pharo is welcome (ML)" <pharo-users(a)lists.pharo.org> > > > Dears, > I'm loading a lot of pictures (70) through the method > "ImageReadWriter>>formFromFileNamed:". > The method is very slow, with images bigger than 5MB. > > Is there a way to improve the pictures speed loading? > > > Thanks > Matteo I think we'll need more details. Here is what I tried: (ImageReadWriter formFromFileNamed: '/Users/sven/Desktop/audio359-main-1600.jpg') asMorph openInWindow. This is a 319K JPG 1600x2263 (Form's Bitmap length is 3620800). [ ImageReadWriter formFromFileNamed: '/Users/sven/Desktop/audio359-main-1600.jpg' ] bench. '2.788 per second' 2 to 3 per second is not very fast, but you create forms of 3.5M - In this case the PluginBasedJPEGReadWriter is used, which should be faster then any of the other formats. What you can do is store a FUEL version of the Form on file, that will be (much) faster (but much larger, it is basically a non-compressed version). Sven

2 2

About balkanisation
by stepharo Nov. 7, 2016

Nov. 7, 2016

Hi I would like that you think a bit about our community and that there is a value in using common tools to share and develop common libraries. Because to me it feels like we are getting balkanize. It may look super cool and be hyper trendy to use github (because like that you can say that you use latest hyper cool features), but I would like to ask especially people building libraries to pay attention that it is important that other people can contribute back easily and that there is an easy way to load/contribute. Today I experienced Bloc - I cannot load code and I cannot contribute. - I saw mdl with a mixture between smalltalkhub and github (sounds super hyper cool) and I saw paul not being able to contribute :( Yes you can say that monticello sucks yes it is terrible yes we all fell like Cobol programmers but at the end of the day. Yes the herb is always greener elsewhere. Yes yes yes. Let us take some facts. We managed pharo and moose with it over the last 8 years successfully and Pharo and moose are not 5 packages together from what I can see. So pay attention about the decision you take. Now we will provide git support (this is 8 months that nicolas is exclusively working/thinking/dreaming about that) and that we are doing experiments (Guille is managing the bootstrap in github). Now when everybody will have its own little project lost on github (I do not count the amount of time I do not find pillar on github because I forget that it is called pillar-markup), what will we do. So we need an infrastructure to handle this and christophe is working on this. I think that you should consider the accidental complexity as something that we can minimise by using patterns and common practices. Now you can think that I'm an idiot and that I have no vision (be my guest) but we should pay attention because we are a small community. Stef

8 22

Iceberg/GitFileTree (was Re: about balkanisation)
by Thierry Goubier Nov. 7, 2016

Nov. 7, 2016

Nicolas, I changed the subject, it seems appropriate. 2016-11-07 16:56 GMT+01:00 Nicolas Passerini <npasserini(a)gmail.com>: > > 2016-11-07 16:26 GMT+01:00 Thierry Goubier <thierry.goubier(a)gmail.com>: > >> >>> The problem I found is that in order to recreate sequential numbers, you >>> have to load all commits into the image. >>> >> >> Why? GitFileTree doesn't need to do that at all! Well, you know the code: >> it just ask git log to rebuild the parent / child relations for a package >> directory, and that is all... Then, like a mcz, the package isn't loaded >> into the image unless you select it in the GUI. So why did you need to load >> all the commits? >> >> Well, at the beginning, I was trying to read all the filetree metadata >> stored in the repository to build the versions as if it was mczs... and >> that was a really bad idea (slow parser for version files, hundred of files >> to scan, many files were even unreadable/broken by conflicts leftovers in >> the filetree repository). >> > > Yes you are right, I expressed myself badly. You don't have to load "the > commits". > But, you still have to create an MCVersionInfo for each version of the > package in the repository. > What you're saying is that you are using the same trick as GitFileTree does, but you have performance issues, which means... we're not talking about exactly the same thing :) (and trust me, my development machine is on the "very" slow end for the Pharo community, so I'm sensitive to performance issues). > > In fact... you can create them, it will not kill you, but I think that > this structure is not so useful, because it is always faster to use > log/revwalk to ask about the relationship between two versions/commits. > Agreed. But you will allways create some sort of MCVersionInfo / MCAncestry to be able to represent the start of this inside your system... Subtypes of MCVersionInfo do work fine there. Just have a nice, correct API call that allow you to compare two versions. (Little trick used in GitFileTree: all ancestry based on a specific commitID is shared among all the versions with it as an ancestor, in memory). Weak version info will take care of cleaning the unneeded stuff anyway. > > Moreover, we do not want to think about packages too much, we prefer to > think in terms of a whole project, so the most interesting concept for us > is the commit and not the version. Therefore, when I build a history > (sometimes I do) I create a graph of commits and not a graph of versions. > As long as you don't track commits for which the package has no version, then that seems perfectly fine. > > Finally, this leads to that, even in the case you want to load a specific > version of a specific pacakge (which is not the normal case for Iceberg), > you would specify a commitish (+ a package name) and not a version name. > I never specify a commitish, I use the GUI or Metacello (and since I'm using baselines, Metacello doesn't uses version numbers except to compare with what is inside the image). GitFileTree only uses commitish to load a version... it just happens to be able to also convert version numbers to a commitish if it is presented with a version number. The attention to version numbering is a thing of people using too much ConfigurationOf, and those people are living in the past. Thierry

1 0

Looking for the library to sort object
by stepharo Nov. 7, 2016

Nov. 7, 2016

Hi I know that "steffen" did a nice library to sort objects and I know that this is a nice extension. But I cannot find it. Stef

3 3

UK Smalltalk User Group Meeting - Monday, November 28th
by Giovanni Corriga Nov. 7, 2016

Nov. 7, 2016

Hi, The next meeting of the UK Smalltalk User Group Meeting will be on Monday, November 28th. We'll meet at our usual venue The Counting House ( http://www.the-counting-house.com/) at 7pm. If you'd like to join us, you can just show up at the pub. You can also sign up in advance at the meeting's Meetup page: https://www.meetup.com/ UKSTUG/events/233683252/ . Thanks, Giovanni

2 1

Pharo Shared Memory with C++ executable
by Dimitris Chloupis Nov. 6, 2016

Nov. 6, 2016

Hey fellow Pharoers I have great news, I was able to finally !!!! understand how shared memory file works and actually do it Why is that great news you may ask. A shared memory file is a file that maps a specific area of memory, which means when you write to a specific part of the memory it writes directly to the file without you having to actually do this. Memory mapped files have the distinct characteristic of being able to share memory between processes. This actually is the fastest way to make two process communicate and exchange data. Its called Inter Process Communication or IPC for short , but is not technically communication because both executables share the memory. This means that its possible to make Pharo use C++ libraries from inside a C++ executable. Because basically the shared memory will allow Pharo and the C++ executables to share data, part of that data can be messages to execute specific functions. What is super cool about shared memory mapped files is not only that they are super fast(its what the OS uses to load shared libraries ie. DLLs) , its not that it automagically stores the shared memory to disk like a Pharo image (so we can store live data) but foremost that a great deal of programming languages can use Memory Mapped Files which basically means they have wrapped the mmap functions of the LibC library (C/C++ Standard Library). That means we can now use this to make Pharo able to use C++ libraries, python libraries and any library from any programming language. Its also a solution when you dont have access to a DLL or making one is very hard like in my case that I use Unreal Engine. As you can imagine this will replace/extend my socket bridge that I made for python. Of course all the above are no magic, you have to have some basic understanding of how C++ works and how memory mapped files work, funny thing is that shared memory is actually simpler to understand than C++. You also have to make the messaging system and manage the synchronisation. But those are details which I will implement anyway. So my challenge right now is to move this to Pharo. I have done it with C++, I made two different executables which share a string with the value "hello". I need two functions to do this one is open, which is the classic way to open files in C/C++ in my case is #define FILEPATH "mmapped.bin" fd = open(FILEPATH, O_RDWR | O_CREAT | O_TRUNC, (mode_t)0600); https://linux.die.net/man/3/open second one is lseek which expand the size of the file to a desired size result = lseek(fd, FILESIZE-1, SEEK_SET); http://man7.org/linux/man-pages/man2/lseek.2.html the third one is mmap which is doing the actually mapping of the shared memory map =(std::string*) mmap(0, FILESIZE, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); http://man7.org/linux/man-pages/man2/mmap.2.html Now my challenges after reading the UFFI documentation are the following 1) How I can pass things like O_CREAT , MAP_SHARED 2) How I cast a void pointer (its what mmap returns) to a C++ string or C++ int In case you want to take a look at the code is part of a project I named Atlas which can be found here https://github.com/kilon/Atlas The file atlas-server.cpp creates the memory mapped file and writes to the shared memory the string "hello" The file atlas-client.cpp gains access to the shared memory and retrieves the "hello" string Atlas basically will be the library that will give Pharo full access to Unreal internals.

2 2

How to consume the MOOC?
by Luke Gorrie Nov. 6, 2016

Nov. 6, 2016

Howdy! Just a question from a confused newbie: What is the easiest way to consume the Pharo MOOC content in English? I have tried to find the content a few times over the past few months but so far I have hit a registration page / short trailer video / video in French. (I keep hoping that a YouTube channel will appear that makes the videos easy to follow in sequence on any device.) Cheers! -Luke

5 5

May I please be added as a contributor to the KevinLanvin/MaterialDesignLite repo on Sthub?
by PAUL DEBRUICKER Nov. 6, 2016

Nov. 6, 2016

This one: http://smalltalkhub.com/#!/~KevinLanvin/MaterialDesignLite/ my username is pdebruic

5 5