- Pharo-users - lists.pharo.org

Have you performed software migration or library update

by Oleksandr Zaytsev

Hello, I am working on a paper about different types of library update (also known as library migration). And I would like to know if you have any experience that I could document. Most modern software depends on many external libraries. And when one of those libraries releases a new version, developers have to update their code according to the changes (e.g. stop using deprecated features, make use of the new functionality, replace old classes with new ones, etc.). Please let me know if you have ever updated your software to a new version of an external library. This may include Pharo libraries (for example, Spec -> Spec2, Roassal2 -> Roassal3, etc.), Java libraries (Maven, Xerces, Log4j, etc.), JavaScript, Python, or any other programming language. I would like to know how much time did it take you to update your code, what kind of changes did you have to make, what problems did you encounter, where did you look for the information on what was changed. I would be very grateful for your help! Oleksandr

April 22, 2021

[ANN] Superluminal v2.1.0 [v2.1.0] released!

by Buenos Aires Smalltalk

April 22, 2021

[ANN] Stargate-Consul v2.1.0 [v2.1.0] released!

by Buenos Aires Smalltalk

April 22, 2021

Re: Seaside disappearing under Windows

by Christopher Fuhrman

I second trying Pharo under WSL for robustness. I wrote a blog entry on how to set it up, and I think it's still valid: https://fuhrmanator.github.io/2019/02/27/Pharo-in-WSL.html - Since you'd be using WSL2, you have to do things differently for the DISPLAY variable: https://github.com/microsoft/WSL/issues/4793#issuecomment-588321333 As to the question of why Windows on Pharo is less stable, I have some experience using over the years other programs ported to Windows from OSS, e.g., LyX and InkScape. I have seen a repeating pattern: key OSS devs of certain libraries are hyper-sensitive to how things are done in Windows (rightly so, some things are very different and require creative solutions that break when Windows changes). We all know OSS resources are limited. Consequently the libraries don't get the love.or testing they need. Despite this I think the Pharo team does a great job supporting Windows. On Thu, 22 Apr 2021 at 15:39, Petter <petter.egesund(a)gmail.com> wrote: > It might be worth trying to run Pharo under wsl if you have to go Windows? > I > do on my job laptop and it works fine for me. > > > > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html > -- Christopher Fuhrman, P.Eng., PhD *Professeur au DÃ©partement de gÃ©nie logiciel et des technologies de l'informationÃTS (Ãcole de technologie supÃ©rieure)* http://profs.etsmtl.ca/cfuhrman +1 514 396 8638 *L'ÃTS est une constituante de l'UniversitÃ© du QuÃ©bec*

April 22, 2021

Comparing source code declarations with #=

by HernÃ¡n Morales Durand

Hi, I want to compare Pharo (8) source code definitions. One source is already "filed in" the image and the other source is in a .st file in chunk format (the result of a file out). I've attached in this mail an example for reproducibility. In my search for a simple equality operator, I found that MCSnapshot(s) understand #= so it could be used like this: | imageSnapshot fileSnapshot | 'MySampleClass.st' asFileReference fileIn. imageSnapshot := (RPackageOrganizer default packageOf: MySampleClass) mcPackage snapshot. fileSnapshot := (MCStReader on: 'MySampleClass.st' asFileReference readStream) snapshot. imageSnapshot = fileSnapshot. But to my surprise, the file snapshot from MCStReader is not reading the class definitions, so comparing them should answer <true> (assuming nothing has been modified) but it always returns <false>. I've also seen that both Ring chunk importers do not read methods. (RingChunkImporter fromStream: 'MySampleClass.st' asFileReference readStream) packages collect: #allMethods. (RGChunkImporter fromStream: 'MySampleClass.st' asFileReference readStream) packages collect: #allMethods. The only "solution" I've found so far is to read into CodeChunk's and then implement the #= CodeImporter chunksFromFileNamed: 'MySampleClass.st'. But I don't like it and also I think there should be a clever way to compare source between image and .st declarations. Any ideas? HernÃ¡n

April 22, 2021

[ANN] Hyperspace v2.6.0 [v2.6.0] released!

by Buenos Aires Smalltalk

April 22, 2021

Re: Seaside disappearing under Windows

by Petter

It might be worth trying to run Pharo under wsl if you have to go Windows? I do on my job laptop and it works fine for me. -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 22, 2021

Re: Segfault with ZnClient and https

by Petter

Hi Sven and thanks for answering! I did try a ldd and all libraries are present. I also started pharo with gdb and it seems problem is somewhere in the combination between libgit2.so, libssl.so and libcrypto.so.1.0.0, most often the crash comes from libssl.so. Exchanging libgit2.so with the os-version make ZnClient work, but then Iceberg on the other hand stops working. I pretty much gave up Fedora and reverted to safe Ubuntu. Maybe I can dig deeper into this later, but we are pretty close to a production setting now :) Petter -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 22, 2021

Re: Seaside disappearing under Windows

by kmo

Hi - Firstly, have you tried asking on the Seaside forum? Might be worth doing. If it were me, I would try to see if it was a problem with my specific application or a a general seaside/windows problem. Does the Seaside demo app work? Can you write a simple Hello World seaside app. Does that work? I would just try to constantly narrow down the problem. Not much help, I know. But these were the thoughts that occurred to me. -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 22, 2021

Re: Seaside disappearing under Windows

by David Pennington

I have looked at the code. There is only one method for reading and one for writing and they both close the file when they have done so thatâs wouldnât ât seem to be the problem. It still doesnât answer why it works on my M1 MacBookAir and not on two Windows boxes :-( > On 20 Apr 2021, at 16:13, Sven Van Caekenberghe <sven(a)stfx.eu> wrote: > > David, > >> On 20 Apr 2021, at 17:01, David Pennington <david(a)totallyobjects.com> wrote: >> >> I have finally given up on trying to sort this out. Under both Windows Server 2012 and Windows 10 my Pharo image just exits leaving nothing, not even a crash file. Hence, I have fixed my MacBookAir so that it stays alive with the lid down and I am now using that as my Seaside server. It has been up for the last 4 days so it seems that this is a Windows only problem and (hopefully) not my code. >> >> In passing, I am a high-level programmer in Pharo in that I do not use any special calls or processes. >> >> Davids >> Totally `Objects > > I believe you are using some form of persistency that you built yourself, with data written in STON to files. Assuming you store each object in its own file, you are then also doing a lot of IO. Are you sure that you are properly closing your file streams, in all circumstances ? > > It is possibly that you run out of IO resources. > > Also, garbage collection on a server image (without a UI) is sometimes a bit strange, it can take a while to kick in. > > I am just guessing here. > > It is still not good that the image/vm quits without any notice, that is for sure. > > Sven

April 22, 2021