- Pharo-users - lists.pharo.org

We haven't had a design debate in a while - thoughts on CD.buy article...

by Tim Mackinnon

Hi guys - someone pointed out this article from a UK colleague which seems to be causing a flurry of discussion - but I always like the insights of this group. I had an instant reaction when I read it - but curious what people here think in 2021. https://www.linkedin.com/posts/jasongorman_the-year-is-2021-and-people-stil… Tim

April 16, 2021

Re: Crash after http request, see the dump file

by David Pennington

Sorry to but in here but I have just moved my Pharo/Seaside server software from my MacBookAir, where it has run without issue for the last couple of weeks, onto firstly a Lenovo desktop running Windows 2012 server and then, in frustration , onto a Dell laptop but it hasnât managed to run for 24 hours without crashing (by crashing, I mean Pharo just disappears!). I donât us any OsProcess, just ordinary Seaside and Smalltalk code. It has happened three days in a row at different times with different pages being accessed. Both Windows machines have ample disk space (SSD) and 16GB memory. How do I get a dump as I canât see any files that would be regarded as so? David > On 15 Apr 2021, at 14:55, Guillermo Polito <guillermopolito(a)gmail.com> wrote: > > Hi Davide, > > Taking a look at the dump, Iâd say it looks like a memory corruption. > It crashes when running the GC, but very probably because the heap was already corrupt. > > > 5 Pharo 0x000000010582ea3b mapStackPages + 345 > 6 Pharo 0x000000010582e0ff mapInterpreterOops + 16 > 7 Pharo 0x000000010580af54 doScavenge + 648 > 8 Pharo 0x000000010581305e scavengingGCTenuringIf + 333 > 9 Pharo 0x00000001057f663a checkForEventsMayContextSwitch + 111 > 10 Pharo 0x00000001057fa815 ceStackOverflow + 110 > 11 ??? 0x000000010b25c62b 0x0 + 4481992235 > 12 Pharo 0x00000001057e2b0a interpret + 632 > > Does this happen often in your setup? > > G > > >> El 14 abr 2021, a las 10:21, Davide Varvello via Pharo-users <pharo-users(a)lists.pharo.org <mailto:pharo-users@lists.pharo.org>> escribiÃ³: >> >> Hi Sven, >> It has not happened today. I also use a lot of OsProcess, but I don't know >> if it is related to it. >> Cheers >> Davide >> >> >> >> Sven Van Caekenberghe-2 wrote >>> I can't see anything networking, HTTP or Zinc related in the dump. >>> It looks like something went wrong during garbage collection. >>> I suppose this is not a repeatable case, is it ? >>> >>>> On 14 Apr 2021, at 09:22, Davide Varvello via Pharo-users < >> >>> pharo-users@.pharo >> >>> > wrote: >>>> >>>> Hi guys, >>>> I'm working on Pharo 8 on BigSur >>>> (Pharo 8.0.0 Build information: >>>> Pharo-8.0.0+build.1128.sha.9f6475d88dda7d83acdeeda794df35d304cf620d (64 >>>> Bit)) >>>> >>>> Yesterday after an http call to my zinc server the image crashed. >>>> Can you please take a look to the dump file and tell me what happened? >>>> >>>> It would be worth so much to me, thank you >>>> >>>> The file is here: https://pastebin.com/SrE6gcCv <https://pastebin.com/SrE6gcCv> >>>> >>>> >>>> Davide >>>> >>>> >>>> >>>> -- >>>> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html <http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html> >> >> >> >> >> >> -- >> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html <http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html>

April 15, 2021

Re: Crash after http request, see the dump file

by Davide Varvello

Hi Guillermo, Nope, it was the first time, but I worry it can happen in the future. Cheers Davide Guillermo Polito wrote > Hi Davide, > > Taking a look at the dump, Iâd say it looks like a memory corruption. > It crashes when running the GC, but very probably because the heap was > already corrupt. > > > 5 Pharo 0x000000010582ea3b mapStackPages + > 345 > 6 Pharo 0x000000010582e0ff > mapInterpreterOops + 16 > 7 Pharo 0x000000010580af54 doScavenge + > 648 > 8 Pharo 0x000000010581305e > scavengingGCTenuringIf + 333 > 9 Pharo 0x00000001057f663a > checkForEventsMayContextSwitch + 111 > 10 Pharo 0x00000001057fa815 ceStackOverflow > + 110 > 11 ??? 0x000000010b25c62b 0x0 + > 4481992235 > 12 Pharo 0x00000001057e2b0a interpret + 632 > > Does this happen often in your setup? > > G > > >> El 14 abr 2021, a las 10:21, Davide Varvello via Pharo-users < > pharo-users@.pharo > > escribiÃ³: >> >> Hi Sven, >> It has not happened today. I also use a lot of OsProcess, but I don't >> know >> if it is related to it. >> Cheers >> Davide >> >> >> >> Sven Van Caekenberghe-2 wrote >>> I can't see anything networking, HTTP or Zinc related in the dump. >>> It looks like something went wrong during garbage collection. >>> I suppose this is not a repeatable case, is it ? >>> >>>> On 14 Apr 2021, at 09:22, Davide Varvello via Pharo-users < >> >>> pharo-users@.pharo >> >>> > wrote: >>>> >>>> Hi guys, >>>> I'm working on Pharo 8 on BigSur >>>> (Pharo 8.0.0 Build information: >>>> Pharo-8.0.0+build.1128.sha.9f6475d88dda7d83acdeeda794df35d304cf620d (64 >>>> Bit)) >>>> >>>> Yesterday after an http call to my zinc server the image crashed. >>>> Can you please take a look to the dump file and tell me what happened? >>>> >>>> It would be worth so much to me, thank you >>>> >>>> The file is here: https://pastebin.com/SrE6gcCv >>>> >>>> >>>> Davide >>>> >>>> >>>> >>>> -- >>>> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html >> >> >> >> >> >> -- >> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html >> <http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html> -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 15, 2021

Re: Crash after http request, see the dump file

by Guillermo Polito

Hi Davide, Taking a look at the dump, Iâd say it looks like a memory corruption. It crashes when running the GC, but very probably because the heap was already corrupt. 5 Pharo 0x000000010582ea3b mapStackPages + 345 6 Pharo 0x000000010582e0ff mapInterpreterOops + 16 7 Pharo 0x000000010580af54 doScavenge + 648 8 Pharo 0x000000010581305e scavengingGCTenuringIf + 333 9 Pharo 0x00000001057f663a checkForEventsMayContextSwitch + 111 10 Pharo 0x00000001057fa815 ceStackOverflow + 110 11 ??? 0x000000010b25c62b 0x0 + 4481992235 12 Pharo 0x00000001057e2b0a interpret + 632 Does this happen often in your setup? G > El 14 abr 2021, a las 10:21, Davide Varvello via Pharo-users <pharo-users(a)lists.pharo.org> escribiÃ³: > > Hi Sven, > It has not happened today. I also use a lot of OsProcess, but I don't know > if it is related to it. > Cheers > Davide > > > > Sven Van Caekenberghe-2 wrote >> I can't see anything networking, HTTP or Zinc related in the dump. >> It looks like something went wrong during garbage collection. >> I suppose this is not a repeatable case, is it ? >> >>> On 14 Apr 2021, at 09:22, Davide Varvello via Pharo-users < > >> pharo-users@.pharo > >> > wrote: >>> >>> Hi guys, >>> I'm working on Pharo 8 on BigSur >>> (Pharo 8.0.0 Build information: >>> Pharo-8.0.0+build.1128.sha.9f6475d88dda7d83acdeeda794df35d304cf620d (64 >>> Bit)) >>> >>> Yesterday after an http call to my zinc server the image crashed. >>> Can you please take a look to the dump file and tell me what happened? >>> >>> It would be worth so much to me, thank you >>> >>> The file is here: https://pastebin.com/SrE6gcCv >>> >>> >>> Davide >>> >>> >>> >>> -- >>> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html > > > > > > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html <http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html>

April 15, 2021

UK Smalltalk User Group meeting - Wednesday April 28th

by Giovanni Corriga

The next meeting of the UK Smalltalk User Group will be on Wednesday, April 28th. Esteban Lorenzano will give us an update on Pharo: What's new? What is updated? What remains? This talk will present (and show working!) the Pharo 9 roadmap and principal changes made to Pharo since the release of the previous version. Esteban studied Computer Sciences at Universidad de Buenos Aires, and worked since 1994 in many object-oriented and low-level technologies in different software companies, serving in various positions from junior programmer to senior architect. In 2007 he co-founded Smallworks to offer Pharo-based agile development projects. Since 2012 he dedicated full time to developing the Pharo code and community. He works for the Pharo Consortium in Lille, France, as lead developer for Pharo and being responsible with the coordination of new releases and the implementation and maintenance of several Pharo libraries.Given the current COVID-19 restrictions, this will be an online meeting from home. If you'd like to join us, please sign up in advance on the meeting's Meetup page ( https://www.meetup.com/UKSTUG/events/cbklbryccgblc/ ) to receive the meeting details. Donât forget to bring your laptop and drinks!

April 15, 2021

Re: Crash after http request, see the dump file

by Davide Varvello

Hi Sven, It has not happened today. I also use a lot of OsProcess, but I don't know if it is related to it. Cheers Davide Sven Van Caekenberghe-2 wrote > I can't see anything networking, HTTP or Zinc related in the dump. > It looks like something went wrong during garbage collection. > I suppose this is not a repeatable case, is it ? > >> On 14 Apr 2021, at 09:22, Davide Varvello via Pharo-users < > pharo-users@.pharo > > wrote: >> >> Hi guys, >> I'm working on Pharo 8 on BigSur >> (Pharo 8.0.0 Build information: >> Pharo-8.0.0+build.1128.sha.9f6475d88dda7d83acdeeda794df35d304cf620d (64 >> Bit)) >> >> Yesterday after an http call to my zinc server the image crashed. >> Can you please take a look to the dump file and tell me what happened? >> >> It would be worth so much to me, thank you >> >> The file is here: https://pastebin.com/SrE6gcCv >> >> >> Davide >> >> >> >> -- >> Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 14, 2021

Re: Crash after http request, see the dump file

by Sven Van Caekenberghe

I can't see anything networking, HTTP or Zinc related in the dump. It looks like something went wrong during garbage collection. I suppose this is not a repeatable case, is it ? > On 14 Apr 2021, at 09:22, Davide Varvello via Pharo-users <pharo-users(a)lists.pharo.org> wrote: > > Hi guys, > I'm working on Pharo 8 on BigSur > (Pharo 8.0.0 Build information: > Pharo-8.0.0+build.1128.sha.9f6475d88dda7d83acdeeda794df35d304cf620d (64 > Bit)) > > Yesterday after an http call to my zinc server the image crashed. > Can you please take a look to the dump file and tell me what happened? > > It would be worth so much to me, thank you > > The file is here: https://pastebin.com/SrE6gcCv > > > Davide > > > > -- > Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 14, 2021

Crash after http request, see the dump file

by Davide Varvello

Hi guys, I'm working on Pharo 8 on BigSur (Pharo 8.0.0 Build information: Pharo-8.0.0+build.1128.sha.9f6475d88dda7d83acdeeda794df35d304cf620d (64 Bit)) Yesterday after an http call to my zinc server the image crashed. Can you please take a look to the dump file and tell me what happened? It would be worth so much to me, thank you The file is here: https://pastebin.com/SrE6gcCv Davide -- Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

April 14, 2021

Re: Whats the easiest/cheapest way to run a Pharo web app in 2021?

by Norbert Hartl

Thanks! What Iâve learned the last 3 decades in IT is that you can decide about the kind of problems you will have not the amount of problems you will have. Trying to delegate complexity turns the kind of problems into complexer problems while the amount stays the same. Norbert > Am 13.04.2021 um 18:15 schrieb jtuchel(a)objektfabrik.de: > > Norbert, > > you nailed it ;-) Easy and cheap and lots of layers of magic infrastructure nailed and hammered onto each other don't go well with each other. I am still a fan of having some server - virtual or physical - that I can reinstall from scratch and as little moving parts as possible on top of that. I rather have a little collection of sheets on which I wrote down each single step of installing what's needed and a good backup of ini files and whatnot than a pile of "stuff" that makes things "easy". > > I know, that's so 90ies, but hey, we're using Smalltalk from the 70ies ;-) > > Joachim > > > > Am 13.04.21 um 18:09 schrieb Norbert Hartl: >> I read the title which contains âeasiest/cheapestâ and then you drop names âdockerâ, "CI which injectsâ and âstages" â¦ this invalidates at least âeasiestâ and maybe âcheapestâ as well. I think I missed the intent of this mail. So can you give a few more details otherwise it is hard to meet the âeasiest on this level of requirementsâ >> >> Norbert >> >>> Am 12.04.2021 um 10:36 schrieb Tim Mackinnon <tim(a)testit.works>: >>> >>> That is cheap 3e/m is definitely worth consideringâ¦ but I guess you do have to take care of your own patching etc right (which isnât necessarily horrible, but does require a bit of extra effort to track things). I was interested in whether the next step up in the food chain using Docker images that are hosted for you, might lessen the burden a bit? E.g. if your CI injects a pharo image into the latest âsafeâ docker image from the community - then hopefully you are insulated from all of this. It does look like this is becoming a reality if that dockerize.io soln plays out (I got that working, its cheap - however they havenât answered any of my email queriesâ¦ so I do wonder how real it actually is). The fallback would definitely be something like Hetzner or Digital Ocean I guess. >>> >>> Tim >>> >>>> On 12 Apr 2021, at 08:47, Norbert Hartl <norbert(a)hartl.name> wrote: >>>> >>>> >>>> >>>>> Am 12.04.2021 um 04:02 schrieb Jeff Gray <jeff(a)rogerthedog.com>: >>>>> >>>>> Considering easiest and cheapest, there's always self hosting, or are you >>>>> discounting that idea? >>>>> Most geeks have a bit of spare hardware laying around and broadband >>>>> up-speeds aren't too bad. >>>>> I'm guessing that if we are in the $5 a month ball park then we aren't >>>>> needing a guaranteed up time. >>>>> >>>> My cloud instance is 3â¬/month. With an additional 20% amount the instance has a backup. And setting it up is way simpler then getting dynamic DNS updates and all of that configured. Times have changed a bit. >>>> >>>> >>>> Norbert > > > -- > ----------------------------------------------------------------------- > Objektfabrik Joachim Tuchel mailto:jtuchel@objektfabrik.de > Fliederweg 1 http://www.objektfabrik.de > D-71640 Ludwigsburg http://joachimtuchel.wordpress.com > Telefon: +49 7141 56 10 86 0 Fax: +49 7141 56 10 86 1 >

April 13, 2021

Re: Whats the easiest/cheapest way to run a Pharo web app in 2021?

by Tim Mackinnon

This raises a good point from Norbert - what does simple mean? I think we all agree on what is cheap - in my mind sub 10euros (ideally 5) - which seems totally doable from what people are saying (of course this goes up if you want big databases and load balanced stuff etc - although you might squeeze it in from the suggestions so far - but lets take this one as given - Iâd like to run some simple apps for me or my family). So revisiting simple - I started out thinking Docker was overkill and Swarms and Kubernetes made me sick (Iâm sure there is a point where you get to these, but then you probably have a bigger team and lots more skills to potentially consider it - although my observation to date was that these things seems to suck up needless time). HOWEVER - I acknowledge that times have moved on, and I was deeply impressed with Pabloâs article on quickly standing up Pharo in a Docker image based on a community containerâ¦ what struck me about this was that I didnât really need to know or care about underlying Linux - and apt-get update and friendsâ¦ nor about sysctrl and restarting thingsâ¦ so for me simplicity is eliminating all the OS, patches and updates and potentially even scaling - and they get handled in a not-Kubernetes way (or at least the K is totally hidden from me). Equally - for persistence, just want a db as a service - with credentials injected into my image so that it can be safely updated/password rotated etc without me caring much. I can just connect to it (and ideally its backed up for me, if I tick a box and pay a bit more money). I think we may be close to this - particularly as Pierce pointed out that you can build your ST image outside of a container (like many do today - and Github/lab have reasonable tooling to let you metacello a clean image and keep logs to see what has happened, as well as cache the right bits for you so its nice and fast) and then you simply ADD it into Docker with a few lines of Dockerfile. So I am wondering if in 2021, whether Docker might actually have got us to a point of simplicity, and for simple workloads if something like dockerize.io <http://dockerize.io/> (or others - but it seems the closest to what I am imagining although Iâve not used it much so far) abstracts you far enough from the Linux/Networking/Secuirty drudgery that makes web development suck. (Donât get me wrong - I know some of you are wizards at this stuff, and its second nature to jump into bash and do the right things - but I find that while I can get so far, inevitably you get reams of console slush where some package conflicts with another and then you are on stackoverflow or here trying to work out how your distro has cocked something up, or you missed some special switch to keep things sane). But I am curious about what others have experienced - as I know many of you have played with Swarm, Kubernetes etc, and equally maybe native, non Docker has a way of avoiding the OS updates and restart magic in a simpler way? I thought DigitalOcean might do this - but it still seemed that you need to go lower level than I have time for. Tim > On 13 Apr 2021, at 17:48, Christian Haider <christian.haider(a)smalltalked-visuals.com> wrote: > > Yes, this is my camp ð. > > Every layer or tool adds many cool features (their raison d'Ãªtre) which you donât need for the "easy" setup. > You typically need only a handful of the configuration parameters at most (imagine PostgreSQL, Apache, Jenkins etc.). > Yet, all the myriad of super-duper feature settings add to confusion and are nice traps for interesting bugs. > > The complexity of a system is probably exponential to the number of tools or layers... > And each one of them wants to be updated... > And each has its own documentation (which you have to read when you get unlucky) with its own conventions... > > No thanks. I try to keep my stack as simple as possible. > Thatâs also a good reason for Smalltalk, because it makes it much easier to keep all pieces together. > > Happy hacking, > Christian > >> -----UrsprÃ¼ngliche Nachricht----- >> Von: jtuchel(a)objektfabrik.de <jtuchel(a)objektfabrik.de> >> Gesendet: Dienstag, 13. April 2021 18:15 >> An: pharo-users(a)lists.pharo.org >> Betreff: [Pharo-users] Re: Whats the easiest/cheapest way to run a Pharo >> web app in 2021? >> >> Norbert, >> >> you nailed it ;-) Easy and cheap and lots of layers of magic infrastructure >> nailed and hammered onto each other don't go well with each other. I am >> still a fan of having some server - virtual or physical >> - that I can reinstall from scratch and as little moving parts as possible on top >> of that. I rather have a little collection of sheets on which I wrote down each >> single step of installing what's needed and a good backup of ini files and >> whatnot than a pile of "stuff" that makes things "easy". >> >> I know, that's so 90ies, but hey, we're using Smalltalk from the 70ies ;-) >> >> Joachim >> >> >> >> Am 13.04.21 um 18:09 schrieb Norbert Hartl: >>> I read the title which contains âeasiest/cheapestâ and then you drop >>> names âdockerâ, "CI which injectsâ and âstages" â¦ this invalidates at >>> least âeasiestâ and maybe âcheapestâ as well. I think I missed the >>> intent of this mail. So can you give a few more details otherwise it >>> is hard to meet the âeasiest on this level of requirementsâ >>> >>> Norbert >>> >>>> Am 12.04.2021 um 10:36 schrieb Tim Mackinnon <tim(a)testit.works>: >>>> >>>> That is cheap 3e/m is definitely worth consideringâ¦ but I guess you do >> have to take care of your own patching etc right (which isnât necessarily >> horrible, but does require a bit of extra effort to track things). I was >> interested in whether the next step up in the food chain using Docker images >> that are hosted for you, might lessen the burden a bit? E.g. if your CI injects a >> pharo image into the latest âsafeâ docker image from the community - then >> hopefully you are insulated from all of this. It does look like this is becoming >> a reality if that dockerize.io soln plays out (I got that working, its cheap - >> however they havenât answered any of my email queriesâ¦ so I do wonder >> how real it actually is). The fallback would definitely be something like >> Hetzner or Digital Ocean I guess. >>>> >>>> Tim >>>> >>>>> On 12 Apr 2021, at 08:47, Norbert Hartl <norbert(a)hartl.name> wrote: >>>>> >>>>> >>>>> >>>>>> Am 12.04.2021 um 04:02 schrieb Jeff Gray <jeff(a)rogerthedog.com>: >>>>>> >>>>>> Considering easiest and cheapest, there's always self hosting, or >>>>>> are you discounting that idea? >>>>>> Most geeks have a bit of spare hardware laying around and broadband >>>>>> up-speeds aren't too bad. >>>>>> I'm guessing that if we are in the $5 a month ball park then we >>>>>> aren't needing a guaranteed up time. >>>>>> >>>>> My cloud instance is 3â¬/month. With an additional 20% amount the >> instance has a backup. And setting it up is way simpler then getting dynamic >> DNS updates and all of that configured. Times have changed a bit. >>>>> >>>>> >>>>> Norbert >> >> >> -- >> ----------------------------------------------------------------------- >> Objektfabrik Joachim Tuchel mailto:jtuchel@objektfabrik.de >> Fliederweg 1 http://www.objektfabrik.de >> D-71640 Ludwigsburg http://joachimtuchel.wordpress.com >> Telefon: +49 7141 56 10 86 0 Fax: +49 7141 56 10 86 1 >

April 13, 2021