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

Re: [Pharo-project] I give up. Squeaksource has lost lot ofversions again

by Bernat Romagosa

I don't understand why is there a thread identical to this one every two or three weeks. This discussion was already held a long time ago and its outcome was that someone was hired to implement a new code repository, Smalltalkhub's the name. This problem is in its way to be permanently solved, let's all stop speculating about how the system will be and how great would it be if it had feature X. I'm sure Smalltalkhub will be _miles_ ahead of Squeaksource and we'll all be able to contribute our ideas once it is up, running and being actively tested. Of course I expect ESUG thought not only of getting it implemented but also of its maintenance, so I propose we just wait for it to come out and bury this fruitless discussion once and for all. In the meantime, I guess we all have to be patient and keep putting up with an obsolete system for a little bit longer. Cheers, Bernat

May 17, 2011

Re: [Pharo-project] I give up. Squeaksource has lost lot ofversions again

by Rob Withers

Why don't you all write a git client in Pharo? Then you could use github as the shared repository. Monticello packages could be the git projects. You could define git sub-modules to support dependencies to other projects. -----Original Message----- From: Igor Stasenko Sent: Tuesday, May 17, 2011 10:19 AM To: Pharo-project(a)lists.gforge.inria.fr Subject: Re: [Pharo-project] I give up. Squeaksource has lost lot ofversions again On 17 May 2011 15:57, <csrabak(a)bol.com.br> wrote: > Em 17/05/2011 10:32, StÃ©phane Ducasse < stephane.ducasse(a)inria.fr > > escreveu: > >> >> Now this is why ESUG payed objectFusion to propose a new >> >> application. >> >> >> > Is there a plan to provide a migration from squeaksource to new >> > system? Because if not, then i'm afraid, most people will keep >> > using SqS and imagine how many configs/wiki pages and other stuff >> > needs to be updated. >> not that I know. I would give a certain amount of time to people >> to migrate (with their own little fingers) put all the mcz on a ftp >> somewhere close squeaksource > > Given you're being assertive about this, I think it's time to add up > to Igor's stance: > >> > I'd prefer to fully replace SqS with new system at some day >> > without need of migrating projects by all developers and updating >> > links/configs etc. > > Expecting a disparate set of developers with different agendas and > different priorities "to migrate (with their own little fingers)" is > putting too much faith the new system would be _sohhh_ _goooood_ that > the masses will feel compelled to. > well comparing to more and more frequent down times, and a data loss.. i think a system which prove to be at least twice stable than current will be a big advantage. But you will never know before you try. And again: who maintaining squeaksource? Nobody. And now ask yourself: how you will manage your risks in such situation? Will you prefer, maybe a bit flaky system, which has a maintainer who will eventually fix all issues, or a system which "lets hope it will works a bit more" and nobody who are maintaining it? I don't expect that new stuff will be _sohhh_ _goooood_ out of the box. What i do expect however, that new system will be actively maintained. Otherwise if we don't have such guarantees then i don't see reason do migrate at all, because couple years later we will be in same situation: having problems but nobody to address them. > my 0.019999... > > -- > Cesar Rabak > -- Best regards, Igor Stasenko AKA sig.

May 17, 2011

Re: [Pharo-project] I give up. Squeaksource has lost lot of versions again

by Igor Stasenko

On 17 May 2011 15:57, <csrabak(a)bol.com.br> wrote: > Em 17/05/2011 10:32, StÃ©phane Ducasse < stephane.ducasse(a)inria.fr > escreveu: > >> >> Now Â this Â is Â why Â ESUG Â payed objectFusion Â to Â propose Â a Â new >> >> application. >> >> >> > Â Is there a Â plan to provide a migration Â from squeaksource to new >> > system? Â Because Â if not, then Â i'm afraid, most people Â will keep >> > using SqS and imagine how Â many configs/wiki pages and other stuff >> > needs to be updated. >> Â not that I Â know. Â I would give a certain amount Â of time to people >> to migrate (with their own little Â fingers) put all the mcz on a ftp >> somewhere close squeaksource > > Given you're being assertive about this, I think it's time to add up > to Igor's stance: > >> > I'd Â prefer to Â fully Â replace SqS Â with Â new system Â at some Â day >> > without need of migrating Â projects by all developers and updating >> > links/configs etc. > > Expecting a disparate set of developers with different agendas and > different priorities "to migrate (with their own little Â fingers)" is > putting too much faith the new system would be _sohhh_ _goooood_ that > the masses will feel compelled to. > well comparing to more and more frequent down times, and a data loss.. i think a system which prove to be at least twice stable than current will be a big advantage. But you will never know before you try. And again: who maintaining squeaksource? Nobody. And now ask yourself: how you will manage your risks in such situation? Will you prefer, maybe a bit flaky system, which has a maintainer who will eventually fix all issues, or a system which "lets hope it will works a bit more" and nobody who are maintaining it? I don't expect that new stuff will be _sohhh_ _goooood_ out of the box. What i do expect however, that new system will be actively maintained. Otherwise if we don't have such guarantees then i don't see reason do migrate at all, because couple years later we will be in same situation: having problems but nobody to address them. > my 0.019999... > > -- > Cesar Rabak > -- Best regards, Igor Stasenko AKA sig.

May 17, 2011

Re: [Pharo-project] Basic tricks for improving a serializer?

by Igor Stasenko

On 17 May 2011 15:51, <csrabak(a)bol.com.br> wrote: > Mariano, > > It is early in the morning right here and I might have read too hastily to your post, but from the diagonal reading I gave it seems you're going quickly to reinvent the wheel! > > Those issues you rise are typical of database (engine) programming. Â Also, IIRC Smalltalk has another approach to persistence of objects called BOSS, have you given a look at it and check if a mix of Fuel + BOSS would accomplish your intent? > I think its implemented from scratch for one big reason: devil is always in details. Object format, streams, protocol and dozen of various things may differ from one smalltalk dialect to another one. Also, i assume most of us are never heard about BOSS. Just googled it and found couple citations which related to VW world, but not to Squeak/Pharo. Is it runs on Pharo? What its license? But by saying that, i'm a of course in favor of having more and more cross-dialect solutions. > -- > Cesar Rabak > > > Em 17/05/2011 10:21, Mariano Martinez Peck < marianopeck(a)gmail.com > escreveu: > Hi guys. Together with Martin we are developing a fast object graph serializer called Fuel. The idea is to use a pickle format similar to VW Parcels. Right now we are able to correctly serialize (almost) all kind of objects (from different classes and so on). The algorithm is working quite well and we are already quite fast. We also know how to encode more or less efficient some of the "primitive" classes like Floats, SmallIntegers, ByteString, ByteArray, etc, etc. > > What we really don't know anything about streams. For example, until now, we were using just a MultiByteFileStream and doing all the time things like #next:Â #nextPutAll:Â #nextWordPut: etc... > > Now Igor told us for example, to use a buffer like this: > > Â | bufferStream | > Â bufferStream := ByteArray new writeStream. > > Â (FLSerializer on: bufferStream) > Â serialize: anArrayToSerialize. > > Â aFileStream nextPutAll: bufferStream contents. > > and that it at least 2 times faster than we were doing.... I guess it is because it goes to the disk only once. But MultiByteFileStream uses a buffer, doesn't it ? > > For reading (deserialization) we are not doing nothing special. We just open the file stream and we read with #nextÂ #nextString ... etc. > > Another thing is which kind of file stream should we use. MultiByteFileStream ? StandardFileStream ? > > So...before spending time in some optimizations, are there "known" things thatÂ we should check? > > Thanks in advance, > > > -- > Mariano > http://marianopeck.wordpress.com > > > > -- Best regards, Igor Stasenko AKA sig.

May 17, 2011

Re: [Pharo-project] Basic tricks for improving a serializer?

by Igor Stasenko

On 17 May 2011 15:51, <csrabak(a)bol.com.br> wrote: > Mariano, > > It is early in the morning right here and I might have read too hastily to your post, but from the diagonal reading I gave it seems you're going quickly to reinvent the wheel! > > Those issues you rise are typical of database (engine) programming. Â Also, IIRC Smalltalk has another approach to persistence of objects called BOSS, have you given a look at it and check if a mix of Fuel + BOSS would accomplish your intent? > I think its implemented from scratch for one big reason: devil is always in details. Object format, streams, protocol and dozen of various things may differ from one smalltalk dialect to another one. Also, i assume most of us are never heard about BOSS. Just googled it and found couple citations which related to VW world, but not to Squeak/Pharo. Is it runs on Pharo? What its license? But by saying that, i'm a of course in favor of having more and more cross-dialect solutions. > -- > Cesar Rabak > > > Em 17/05/2011 10:21, Mariano Martinez Peck < marianopeck(a)gmail.com > escreveu: > Hi guys. Together with Martin we are developing a fast object graph serializer called Fuel. The idea is to use a pickle format similar to VW Parcels. Right now we are able to correctly serialize (almost) all kind of objects (from different classes and so on). The algorithm is working quite well and we are already quite fast. We also know how to encode more or less efficient some of the "primitive" classes like Floats, SmallIntegers, ByteString, ByteArray, etc, etc. > > What we really don't know anything about streams. For example, until now, we were using just a MultiByteFileStream and doing all the time things like #next:Â #nextPutAll:Â #nextWordPut: etc... > > Now Igor told us for example, to use a buffer like this: > > Â | bufferStream | > Â bufferStream := ByteArray new writeStream. > > Â (FLSerializer on: bufferStream) > Â serialize: anArrayToSerialize. > > Â aFileStream nextPutAll: bufferStream contents. > > and that it at least 2 times faster than we were doing.... I guess it is because it goes to the disk only once. But MultiByteFileStream uses a buffer, doesn't it ? > > For reading (deserialization) we are not doing nothing special. We just open the file stream and we read with #nextÂ #nextString ... etc. > > Another thing is which kind of file stream should we use. MultiByteFileStream ? StandardFileStream ? > > So...before spending time in some optimizations, are there "known" things thatÂ we should check? > > Thanks in advance, > > > -- > Mariano > http://marianopeck.wordpress.com > > > > -- Best regards, Igor Stasenko AKA sig.

May 17, 2011

Re: [Pharo-project] I give up. Squeaksource has lost lot of versions again

by csrabak＠bol.com.br

Em 17/05/2011 10:32, StÃ©phane Ducasse < stephane.ducasse(a)inria.fr > escreveu: > >> Now this is why ESUG payed objectFusion to propose a new > >> application. > >> > > Is there a plan to provide a migration from squeaksource to new > > system? Because if not, then i'm afraid, most people will keep > > using SqS and imagine how many configs/wiki pages and other stuff > > needs to be updated. > not that I know. I would give a certain amount of time to people > to migrate (with their own little fingers) put all the mcz on a ftp > somewhere close squeaksource Given you're being assertive about this, I think it's time to add up to Igor's stance: > > I'd prefer to fully replace SqS with new system at some day > > without need of migrating projects by all developers and updating > > links/configs etc. Expecting a disparate set of developers with different agendas and different priorities "to migrate (with their own little fingers)" is putting too much faith the new system would be _sohhh_ _goooood_ that the masses will feel compelled to. my 0.019999... -- Cesar Rabak

May 17, 2011

Re: [Pharo-project] Basic tricks for improving a serializer?

by csrabak＠bol.com.br

Mariano, It is early in the morning right here and I might have read too hastily to your post, but from the diagonal reading I gave it seems you're going quickly to reinvent the wheel! Those issues you rise are typical of database (engine) programming. Also, IIRC Smalltalk has another approach to persistence of objects called BOSS, have you given a look at it and check if a mix of Fuel + BOSS would accomplish your intent? -- Cesar Rabak Em 17/05/2011 10:21, Mariano Martinez Peck < marianopeck(a)gmail.com > escreveu: Hi guys. Together with Martin we are developing a fast object graph serializer called Fuel. The idea is to use a pickle format similar to VW Parcels. Right now we are able to correctly serialize (almost) all kind of objects (from different classes and so on). The algorithm is working quite well and we are already quite fast. We also know how to encode more or less efficient some of the "primitive" classes like Floats, SmallIntegers, ByteString, ByteArray, etc, etc.Â What we really don't know anything about streams. For example, until now, we were using just a MultiByteFileStream and doing all the time things like #next:Â #nextPutAll:Â #nextWordPut: etc... Now Igor told us for example, to use a buffer like this: | bufferStream | bufferStream := ByteArray new writeStream. (FLSerializer on: bufferStream) serialize: anArrayToSerialize. aFileStream nextPutAll: bufferStream contents. and that it at least 2 times faster than we were doing.... I guess it is because it goes to the disk only once. But MultiByteFileStream uses a buffer, doesn't it ? For reading (deserialization) we are not doing nothing special. We just open the file stream and we read with #nextÂ #nextString ... etc. Another thing is which kind of file stream should we use. MultiByteFileStream ? StandardFileStream ? So...before spending time in some optimizations, are there "known" things thatÂ we should check? Thanks in advance, -- Mariano http://marianopeck.wordpress.com

May 17, 2011

Re: [Pharo-project] I give up. Squeaksource has lost lot of versions again

by Igor Stasenko

On 17 May 2011 15:37, Philippe Marschall <philippe.marschall(a)netcetera.ch> wrote: > On 05/17/2011 09:53 AM, Igor Stasenko wrote: >> On 16 May 2011 19:13, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: >>> mariano >>> >>> There was a crash and we spend some couple of hours with marcus to see what we can >>> do. Now apparently the applications does not find some of the files that are on the filesystem. >>> So the code is not lost but squeaksource does not see them. >>> >>> This has nothing with Pharo and Smalltalk. Just that Squeaksource was created in 2002 >>> and that it suffers from the success of Squeak and Pharo. >>> >> >> Indeed. It was not designed to be used in such big scale as today. So, >> it is not a bug, it is a feature (or lack of ;) >> And as any other software it requires maintenance and upgrades to meet >> today's requirements. >> >>> Now this is why ESUG payed objectFusion to propose a new application. >>> >> >> Is there a plan to provide a migration from squeaksource to new system? >> Because if not, then i'm afraid, most people will keep using SqS >> and imagine how many configs/wiki pages and other stuff needs to be updated. >> I'd prefer to fully replace SqS with new system at some day without >> need of migrating projects by all developers and updating >> links/configs etc. > > If you don't care about metadata it should be pretty easy to come up a > shell script that downloads all versions of a project from SqueakSource > and uploads when to a new place. > Yes, but the problem is not just about files. SqS is a major part of our community infrastructure. So, if you tell people: here another novel system which you can use to host your projects, how many people you will expect to migrate there? If project aims to replace squeaksource, then i'd prefer to do that literally: replacing it for good of all users of current squeaksource. Otherwise, some projects will be still bound to old infrastructure , and apparently we will never be able to get rid of that 2002 old image, and will be forced to still use it for projects whose author(s)/maintainers decided to not bother migrating on new server. And again, imagine the amount of work migrating a project from one hosting to another: - update configurations with proper urls, update docs and references.. this is a lot of work. And apparently because of that i would not expect that everyone will start migrating their stuf onto new server. Take in account a human factors, like making mistakes etc etc.. so it will be a whole mess. What i propose instead is that one day, people go to squeaksource.com and magically see different frontend (which using different backend), but all data from squeaksource is there and all projects and users etc etc. This will be much more helpful to community if we do it like that. > Cheers > Philippe > -- Best regards, Igor Stasenko AKA sig.

May 17, 2011

Re: [Pharo-project] I give up. Squeaksource has lost lot of versions again

by csrabak＠bol.com.br

Em 16/05/2011 14:13, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> escreveu: > mariano > There was a crash and we spend some couple of hours with marcus to > see what we can do. Now apparently the applications does not find > some of the files that are on the filesystem. So the code is not > lost but squeaksource does not see them. > This has nothing with Pharo and Smalltalk. Just that Squeaksource > was created in 2002 and that it suffers from the success of Squeak > and Pharo. > Now this is why ESUG payed objectFusion to propose a new > application. > > Now again if you give me one engineer full time I'm sure that we > can do that. Do you think that svn or git where created from the > air without sponsoring......? Just curious. Stef, Incurring on the risk of sounding like a cat in the Birds of a Feather, I think the attempt to compare svn or git efforts to SqS (or the whole Monticello infra, for the matter) is not correct, because in former the solutions span a wider "audience" with more than a technology/platform/language, whereas the latter is Squeak/Pharo specific. In Smalltalk we have for each flavor a different version control system! To avoid misunderstandings and/or a stretching of this thread without necessity, let assure you folks I _do_ know the aformentioned systems are "file based" and Smalltalk is different, etc., but my observation is /orthogonal/ to this. my 0.0199999... -- Cesar Rabak

May 17, 2011

Re: [Pharo-project] I give up. Squeaksource has lost lot of versions again

by Philippe Marschall

On 05/17/2011 09:53 AM, Igor Stasenko wrote: > On 16 May 2011 19:13, StÃ©phane Ducasse <stephane.ducasse(a)inria.fr> wrote: >> mariano >> >> There was a crash and we spend some couple of hours with marcus to see what we can >> do. Now apparently the applications does not find some of the files that are on the filesystem. >> So the code is not lost but squeaksource does not see them. >> >> This has nothing with Pharo and Smalltalk. Just that Squeaksource was created in 2002 >> and that it suffers from the success of Squeak and Pharo. >> > > Indeed. It was not designed to be used in such big scale as today. So, > it is not a bug, it is a feature (or lack of ;) > And as any other software it requires maintenance and upgrades to meet > today's requirements. > >> Now this is why ESUG payed objectFusion to propose a new application. >> > > Is there a plan to provide a migration from squeaksource to new system? > Because if not, then i'm afraid, most people will keep using SqS > and imagine how many configs/wiki pages and other stuff needs to be updated. > I'd prefer to fully replace SqS with new system at some day without > need of migrating projects by all developers and updating > links/configs etc. If you don't care about metadata it should be pretty easy to come up a shell script that downloads all versions of a project from SqueakSource and uploads when to a new place. Cheers Philippe

May 17, 2011