May 2011 - Pharo-dev - lists.pharo.org

Re: [Pharo-project] Nile tests red - Pharo 1.3

by Mariano Martinez Peck

On Fri, May 6, 2011 at 10:44 AM, Alain rastoul <Alr.dev(a)free.fr> wrote: > I used odbc-ffi and they are serious problem with the single threaded vm, > you can lock the vm with only 2 connections. > In fact with only one. If the query is long, and ODBC doesn't answer until the query is finished...you are screw :( > That's what I want to check with dbx. I m also interested in glorp. > . Sqlite is good for standalone or small apps. > I'm talking of small to medium business systems from 1-10 users (small) to > 500 (medium) computer users - where customers have dba and other > admistrators who impose things like sql server or oracle. > Yes, exactly. That was the reason behind DBX. MSSQL and Oracle are supported :) > Postgresql could be also a player - don't know. > I feel that pharo could take a place here with seaside apps > > +1 -- Mariano http://marianopeck.wordpress.com

May 6, 2011

Re: [Pharo-project] Nile tests red - Pharo 1.3

by StÃ©phane Ducasse

On May 6, 2011, at 10:44 AM, Alain rastoul wrote: > I used odbc-ffi and they are serious problem with the single threaded vm, you can lock the vm with only 2 connections. > That's what I want to check with dbx. I m also interested in glorp. I would really like to have a good layer build on glorp to support easy CRUDE in Smalltalk The problem of Glorp was that it is a bit heavyweight. > . Sqlite is good for standalone or small apps. > I'm talking of small to medium business systems from 1-10 users (small) to 500 (medium) computer users - where customers have dba and other admistrators who impose things like sql server or oracle. > Postgresql could be also a player - don't know. > I feel that pharo could take a place here with seaside apps > > -- > Alain_rastoul > >

May 6, 2011

Re: [Pharo-project] Popularity of Smalltalk in Software Industry

by Stefan Marr

On 06 May 2011, at 10:30, StÃ©phane Ducasse wrote: > Igor > > We should face it: the FFI and Alien... are not that simple. Laurent spent time to try to cnnect to libs and it did > not work. So we should really improve that aspect. > Igor not everbody is able to code in assembler for nativeboost. > So NB is probably a good insfrastructure but not ready for consumption for everybody. > The interaction with C should be improved :) > > Stef No, Stef, you didn't get it! It is documentation which needs to be improved in the first place. Documentation and its presentation. You know another reason why people love PHP? It is an ugly and awful language, but absolutely everything you ever want to do is covered in the docs, because people actually add examples and comments. And frankly, you book efforts are certainly interesting, but far from being as visible and as accessible as online docs of other languages. That's what Smalltalk sucks at, accessible online documentation. People do google when they have a problem. If it does not pop up on the first page, it does not exists. My impression of Smalltalk documentation is, it is non-existing. I usually do not even care to google, because it is unlikely that I find something. Instead I ask people. (And if you look at the traffic on the mailing list, thats not just me) That's a community issue, a problem of culture. I would never think about asking anyone if I have a problem with Python or PHP. I know everything is just a few google hits away. Ah, and tooling of course... -- Stefan Marr Software Languages Lab Vrije Universiteit Brussel Pleinlaan 2 / B-1050 Brussels / Belgium http://soft.vub.ac.be/~smarr Phone: +32 2 629 2974 Fax: +32 2 629 3525

May 6, 2011

Re: [Pharo-project] Popularity of Smalltalk in Software Industry

by Toon Verwaest

Exactly right. BUT! Very cool job. I'll have a look at it in detail once I start interfacing more thoroughly with real C apps outside of my own natives :) I wonder if being able to extend the object model of Smalltalk for native object wrappers would give you benefit... I guess it would? On 05/06/2011 10:30 AM, StÃ©phane Ducasse wrote: > Igor > > We should face it: the FFI and Alien... are not that simple. Laurent spent time to try to cnnect to libs and it did > not work. So we should really improve that aspect. > Igor not everbody is able to code in assembler for nativeboost. > So NB is probably a good insfrastructure but not ready for consumption for everybody. > The interaction with C should be improved :) > > Stef > > >> So, read about it. It doesn't bites. :) >> >>> As I said, I am totally unaware of the limitations of the system, but that >>> it did what I need, and would have liked to do that from Pharo/Squeak. I >>> know Dolphin has/had certain capabilities. But I don't prefer to use >>> non-open source software if at all possible for development. I also am very >>> preferential towards cross-platform software. Dolphin fails on all accounts. >>> I would choose my Python/Pharo blend over Dolphin any day. I know it >>> introduces some pain, but I am willing to accept the pain. I like tools that >>> allow me to use them where ever I am and whatever I am doing. >>> >>> However, that said, when I look at the facilities Python offers for such >>> capabilities it on appearances looks quite impressive. >>> >>> http://docs.python.org/library/ctypes.html >>> >>> """ >>> ctypes is a foreign function library for Python. It provides C compatible >>> data types, and allows calling functions in DLLs or shared libraries. It can >>> be used to wrap these libraries in pure Python. >>> >>> 15.18.1.1. Loading dynamic link libraries >>> >>> ctypes exports the cdll, and on Windows windll and oledll objects, for >>> loading dynamic link libraries. >>> >>> You load libraries by accessing them as attributes of these objects. cdll >>> loads libraries which export functions using the standard cdecl calling >>> convention, while windll libraries call functions using the stdcall calling >>> convention. oledll also uses the stdcall calling convention, and assumes the >>> functions return a Windows HRESULT error code. The error code is used to >>> automatically raise a WindowsError exception when the function call fails. >>> >>> Here are some examples for Windows. Note that msvcrt is the MS standard C >>> library containing most standard C functions, and uses the cdecl calling >>> convention: >>> >>>>>> from ctypes import * >>>>>> print windll.kernel32 # doctest: +WINDOWS >>> <WinDLL 'kernel32', handle ... at ...> >>>>>> print cdll.msvcrt # doctest: +WINDOWS >>> <CDLL 'msvcrt', handle ... at ...> >>>>>> libc = cdll.msvcrt # doctest: +WINDOWS >> So? Why do you think that you cannot do the same in Pharo? >> Have you digested what is available before doing it in Python? >> >> >>> """ >>> >>> It has examples for windows and linux. >>> >> I feel that you are seriously under-informed. >> Just try googling for: >> >> Squeak + FFI >> Alien >> >> NativeBoost >> >> and just check what you can do using it: >> >> http://www.squeaksource.com/NBOpenGL/ >> >> also check mail archives. >> >> Forgive me, but i really can't understand , why you can't just look >> for what you need by yourself? >> Are google.com banned by your ISP? >> >> For me this subject is interested in aspect, why this information >> (while being available openly) didn't catch your >> attention , so you using python instead. >> (Okay, you might miss some functionality like being able to >> automatically generate bindings to COM interfaces). >> But again, i am sure that if you surf the net or ask right questions >> on mailing list, you will probably discover that >> there is already work being done in that direction. >> And at last, if you feel that there is missing some key functionality >> which you want, and you expect that it is also >> could be useful to community, you can always implement it and share >> with people (instead of using Python ;). >> So, next who will come after you won't find himself in a desert with >> couple of oasises few hundred miles away. >> >>> I didn't read it all as I am unqualified to assess its capabilities or >>> limitations, nor do I presently need it as the win32 extensions do easily >>> and well what I presently need to do. >>> >>> Hopefully it can inform us as to what is expected and doable in alternative >>> languages which are dynamic like Smalltalk. >>> >>> Jimmie >>> >> >> >> -- >> Best regards, >> Igor Stasenko AKA sig. >> >

May 6, 2011

Re: [Pharo-project] Nile tests red - Pharo 1.3

by Alain rastoul

I used odbc-ffi and they are serious problem with the single threaded vm, you can lock the vm with only 2 connections. That's what I want to check with dbx. I m also interested in glorp. . Sqlite is good for standalone or small apps. I'm talking of small to medium business systems from 1-10 users (small) to 500 (medium) computer users - where customers have dba and other admistrators who impose things like sql server or oracle. Postgresql could be also a player - don't know. I feel that pharo could take a place here with seaside apps -- Alain_rastoul

May 6, 2011

Re: [Pharo-project] Making cr to autocomplete as well as tab (in Pharo 1.3)

by Tudor Girba

+1 Doru On 6 May 2011, at 10:06, Johan Brichau wrote: > > On 06 May 2011, at 04:59, Jimmie Houchin wrote: > >> As I stated in my reply I like the idea of using CR. But you do raise a valid point and demonstrate a valid situation. > > When CR was in there before, it was absolutely irritating. > I had to put a space before hitting return just to prevent a completion. In my case, it always happened when typing the method header. > >> My question is... >> Why isn't OCompletion offering the shortest options first? > > OCompletion's ordering is based on your recent activity. See http://www.inf.unisi.ch/phd/robbes/papers/ASE2008-completion.pdf > In my opinion, this works really great. > >> If it did so, then your situation wouldn't occur. >> tree would be selected before treeLayout. > > no, because the situation also occurs when you are typing a new message that does not exist yet. > > If we want to improve completion, I think we should take a look at XCode's completion. It works *very* nice for completion of keyworded message syntax. > > Johan -- www.tudorgirba.com "The coherence of a trip is given by the clearness of the goal."

May 6, 2011

Re: [Pharo-project] Popularity of Smalltalk in Software Industry

by StÃ©phane Ducasse

Igor We should face it: the FFI and Alien... are not that simple. Laurent spent time to try to cnnect to libs and it did not work. So we should really improve that aspect. Igor not everbody is able to code in assembler for nativeboost. So NB is probably a good insfrastructure but not ready for consumption for everybody. The interaction with C should be improved :) Stef > So, read about it. It doesn't bites. :) > >> As I said, I am totally unaware of the limitations of the system, but that >> it did what I need, and would have liked to do that from Pharo/Squeak. I >> know Dolphin has/had certain capabilities. But I don't prefer to use >> non-open source software if at all possible for development. I also am very >> preferential towards cross-platform software. Dolphin fails on all accounts. >> I would choose my Python/Pharo blend over Dolphin any day. I know it >> introduces some pain, but I am willing to accept the pain. I like tools that >> allow me to use them where ever I am and whatever I am doing. >> >> However, that said, when I look at the facilities Python offers for such >> capabilities it on appearances looks quite impressive. >> >> http://docs.python.org/library/ctypes.html >> >> """ >> ctypes is a foreign function library for Python. It provides C compatible >> data types, and allows calling functions in DLLs or shared libraries. It can >> be used to wrap these libraries in pure Python. >> >> 15.18.1.1. Loading dynamic link libraries >> >> ctypes exports the cdll, and on Windows windll and oledll objects, for >> loading dynamic link libraries. >> >> You load libraries by accessing them as attributes of these objects. cdll >> loads libraries which export functions using the standard cdecl calling >> convention, while windll libraries call functions using the stdcall calling >> convention. oledll also uses the stdcall calling convention, and assumes the >> functions return a Windows HRESULT error code. The error code is used to >> automatically raise a WindowsError exception when the function call fails. >> >> Here are some examples for Windows. Note that msvcrt is the MS standard C >> library containing most standard C functions, and uses the cdecl calling >> convention: >> >>>>> from ctypes import * >>>>> print windll.kernel32 # doctest: +WINDOWS >> <WinDLL 'kernel32', handle ... at ...> >>>>> print cdll.msvcrt # doctest: +WINDOWS >> <CDLL 'msvcrt', handle ... at ...> >>>>> libc = cdll.msvcrt # doctest: +WINDOWS >>>> >> > > So? Why do you think that you cannot do the same in Pharo? > Have you digested what is available before doing it in Python? > > >> """ >> >> It has examples for windows and linux. >> > > I feel that you are seriously under-informed. > Just try googling for: > > Squeak + FFI > Alien > > NativeBoost > > and just check what you can do using it: > > http://www.squeaksource.com/NBOpenGL/ > > also check mail archives. > > Forgive me, but i really can't understand , why you can't just look > for what you need by yourself? > Are google.com banned by your ISP? > > For me this subject is interested in aspect, why this information > (while being available openly) didn't catch your > attention , so you using python instead. > (Okay, you might miss some functionality like being able to > automatically generate bindings to COM interfaces). > But again, i am sure that if you surf the net or ask right questions > on mailing list, you will probably discover that > there is already work being done in that direction. > And at last, if you feel that there is missing some key functionality > which you want, and you expect that it is also > could be useful to community, you can always implement it and share > with people (instead of using Python ;). > So, next who will come after you won't find himself in a desert with > couple of oasises few hundred miles away. > >> I didn't read it all as I am unqualified to assess its capabilities or >> limitations, nor do I presently need it as the win32 extensions do easily >> and well what I presently need to do. >> >> Hopefully it can inform us as to what is expected and doable in alternative >> languages which are dynamic like Smalltalk. >> >> Jimmie >> > > > > -- > Best regards, > Igor Stasenko AKA sig. >

May 6, 2011

Re: [Pharo-project] Nile tests red - Pharo 1.3

by StÃ©phane Ducasse

On May 6, 2011, at 12:44 AM, Alain rastoul wrote: > Nice vision. I want it now :) > I think smalltalk is the best choice for that - you know better than me of course > However to make a living with it, it has to interface with other IT systems and standards, so that developers can make an sell a part of the system. Yes > Pharo has seaside-web on one end but may be something is missing with relational databases. They are today the heart of business information systems (I have to try dbx and glorp , odbc is not usable because of locks) Yes this is why we need a good FFI not locking. > Side note, I often wondered myself why object databases didn't catch? Technical issues? Marketing? I talked with somebody that was and is in the business and she told me that RDB nicely killed all the start up on the 80/90e around OODB. Now we should try to get pharo running on gemstone one of these days :). > Btw, pharo has taken a good direction. Nice work. I'll try and see if I can help in this field (databases) which is one of my skill Yes we are really interesting in that! > Cheers > Alain > > -- > Alain_rastoul > >

May 6, 2011

Re: [Pharo-project] Preferred way of removing slice from Inbox

by Mariano Martinez Peck

On Fri, May 6, 2011 at 10:02 AM, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr > wrote: > > On May 6, 2011, at 12:23 AM, Mariano Martinez Peck wrote: > > > > > > > On Thu, May 5, 2011 at 10:22 PM, StÃ©phane Ducasse < > stephane.ducasse(a)inria.fr> wrote: > > > > On May 5, 2011, at 9:00 PM, laurent laffont wrote: > > > > > Hi, > > > > > > Once a SLICE is integrated in PharoCore, how do you remove it from > PharoInbox ? > > > > via boring click click click squeaksource interface > > > > > > mmmmmm shouldn't we be able to create a simple Gofer script that do that? > > yes this could be a good idea. > > Yes, but after sending the mail...I thought...how does Gofer can know WHICH slices from inbox to move ? there is no way. All slices or non. And I think we can never move ALL. We always move only those who has been integrated, but how can we tell that to Gofer? because writing a Gofer script but listing manually each of the slices is even worst than using directly the MonticelloBrowser. cheers > Stef > > > > > lukas? > > > > > > -- > > Mariano > > http://marianopeck.wordpress.com > > > > > -- Mariano http://marianopeck.wordpress.com

May 6, 2011

Re: [Pharo-project] Making cr to autocomplete as well as tab (in Pharo 1.3)

by Johan Brichau

On 06 May 2011, at 04:59, Jimmie Houchin wrote: > As I stated in my reply I like the idea of using CR. But you do raise a valid point and demonstrate a valid situation. When CR was in there before, it was absolutely irritating. I had to put a space before hitting return just to prevent a completion. In my case, it always happened when typing the method header. > My question is... > Why isn't OCompletion offering the shortest options first? OCompletion's ordering is based on your recent activity. See http://www.inf.unisi.ch/phd/robbes/papers/ASE2008-completion.pdf In my opinion, this works really great. > If it did so, then your situation wouldn't occur. > tree would be selected before treeLayout. no, because the situation also occurs when you are typing a new message that does not exist yet. If we want to improve completion, I think we should take a look at XCode's completion. It works *very* nice for completion of keyworded message syntax. Johan

May 6, 2011