Hi all This is the week-end and we worked super well yesterday during the sprint. Lot of good enhancements - Thanks a lot to all the participants. I not really happy to be forced to do it on a sunny saturday but I’m doing it to clarify points. Esteban sent me this text that was posted on Squeak-Dev (I personally do not read squeak related forums because I have not the time and my focus is Pharo, its consortium, my team, my research and my family). We have to react because - We do not really at ***all** understand this email - We did not kicked anybody from our mailing-list from ages - so ron is lying. In the past we even had discussion with ron - so we do not really understand. May be we got problem to log on our mailing-lists. We have no idea because we are working and not looking at such things. - When we migrated smalltalkhub to readonly we payed attention to make sure that private projects stay private. We did not migrated smalltalkhub for fun. We MUST do it or it will be done by our infrastructure! - Now the cryptography packages are MIT and they are public anyway. So again we do not understand anything. We do not get why Ron contacted us because we announced the migration publicly way in advance and we will keep the Smalltalkhub frozen repo for at least next 5 years. I feel really sorry to hear such kind of email because we do not want to fight with anybody. Our goal is to make sure that people can work with Pharo and expand their business and knowledge. We are working hard to make sure that people can invent their future with Pharo and people that know us personally know that we are not lying. S
-------------------------------------------- Stéphane Ducasse 03 59 35 87 52 Assistant: Aurore Dalle FAX 03 59 57 78 50 TEL 03 59 35 86 16 S. Ducasse - Inria 40, avenue Halley, Parc Scientifique de la Haute Borne, Bât.A, Park Plaza Villeneuve d'Ascq 59650 France |
Yes, let us coordinate what to respond. To me this doesn't even sound like Ron. Usually he is a rather quite and reasonable guy. Seems to be a lot of misunderstandings here (again).
As I was the one copying the Cryptography package to github it concerns me a bit. I've been remembered now about yet another strange law suit in the US and indeed we need to raise a bit of awareness for this. So let's figure out what has to be done. Norbert
|
He’s on discord on squeak server. Maybe chatting with him in private could help resolve the misunderstanding. Cheers, Cedrick Le 30 mai 2020 à 18:26, Norbert Hartl <[hidden email]> a écrit :
|
In reply to this post by NorbertHartl
Hi Norbert,
The EFF gives a nice overview of the US situation: https://www.eff.org/deeplinks/2019/08/us-export-controls-and-published-encryption-source-code-explained Although you are not a US citizen, and don't live here, so I don't know how it could affect you. In the link above the EFF does say they provide free consultations about this topic so I'm sure someone there will clear things up for you. Paul NorbertHartl wrote > Yes, let us coordinate what to respond. To me this doesn't even sound like > Ron. Usually he is a rather quite and reasonable guy. Seems to be a lot of > misunderstandings here (again). > As I was the one copying the Cryptography package to github it concerns me > a bit. I've been remembered now about yet another strange law suit in the > US and indeed we need to raise a bit of awareness for this. > > So let's figure out what has to be done. > > Norbert > > >> Am 30.05.2020 um 14:43 schrieb Stéphane Ducasse < > stephane.ducasse@ > >: >> >> Hi all >> >> This is the week-end and we worked super well yesterday during the >> sprint. Lot of good enhancements - Thanks a lot to all the participants. >> I not really happy to be forced to do it on a sunny saturday but I’m >> doing it to clarify points. >> >> Esteban sent me this text that was posted on Squeak-Dev (I personally do >> not read squeak related forums because >> I have not the time and my focus is Pharo, its consortium, my team, my >> research and my family). >> >> We have to react because >> - We do not really at ***all** understand this email >> - We did not kicked anybody from our mailing-list from ages - so ron is >> lying. In the past we even had discussion with ron - so we do not >> really understand. May be we got problem to log on our mailing-lists. >> We have no idea because we are working and not looking at such things. >> - When we migrated smalltalkhub to readonly we payed attention to make >> sure that private projects stay private. >> We did not migrated smalltalkhub for fun. We MUST do it or it will be >> done by our infrastructure! >> - Now the cryptography packages are MIT and they are public anyway. So >> again we do not understand anything. >> >> We do not get why Ron contacted us because we announced the migration >> publicly way in advance and we will keep >> the Smalltalkhub frozen repo for at least next 5 years. >> >> I feel really sorry to hear such kind of email because we do not want to >> fight with anybody. >> Our goal is to make sure that people can work with Pharo and expand their >> business and knowledge. >> We are working hard to make sure that people can invent their future with >> Pharo and people that know us personally >> know that we are not lying. >> >> S >> >> >>> Hi all, >>> >>> I've tried to work with the Pharo group but they keep kicking me out of >>> their mailing list. I've already mentioned this a number of times to >>> the Pharo group but nobody seems to care. >>> >>> BOLD BOLD BOLD PLEASE TAKE THIS SERIOUSLY BOLD BOLD BOLD >>> >>> I am not a lawyer but we used very good lawyers to make the squeaksource >>> repository a safe place to do cryptography work. If you are working on >>> cryptography DO NOT POST your code anywhere except squeaksource. >>> Especially if you are in the USA. The ONLY repository that is approved >>> to host our cryptography code in the USA and therefore not subject to >>> criminal violations is squeaksource. It is a CRIME in the USA to move >>> code and make it available on the internet for everyone to download! It >>> must be hosted on squeaksoruce.com <http://squeaksoruce.com/> or >>> another location that is also properly registered. >>> >>> IF YOU COPIED CRYPTOGRAPHY CODE TO ANOTHER REPOSITORY THAT IS NOT >>> REGISTERED I would recommend you delete it immediately. >>> >>> END BOLD! >>> >>> Please feel free to post this to the Pharo mailing list because they >>> apparently do not want to hear from me! >>> >>> All the best, >>> >>> Ron Teitelbaum >> >> >> -------------------------------------------- >> Stéphane Ducasse >> http://stephane.ducasse.free.fr <http://stephane.ducasse.free.fr/> >> / http://www.pharo.org <http://www.pharo.org/> >> 03 59 35 87 52 >> Assistant: Aurore Dalle >> FAX 03 59 57 78 50 >> TEL 03 59 35 86 16 >> S. Ducasse - Inria >> 40, avenue Halley, >> Parc Scientifique de la Haute Borne, Bât.A, Park Plaza >> Villeneuve d'Ascq 59650 >> France >> -- Sent from: http://forum.world.st/Pharo-Smalltalk-Developers-f1294837.html |
In reply to this post by Stéphane Ducasse
Yes, also the project is MIT so… anyone has the right to take it and use it/modify it/etc. with the only constraint of mention all the contributors.
But anyway, for me the most important question is: How far are we to provide good FFI bindings to a security library (like openssl or/and others) that will allow us to escape this situation? I really prefer not to have a hard to maintain solution that will also annoy the original authors who put its code in MIT but didn’t realise the implications of it. Esteban > On 1 Jun 2020, at 08:59, Sven Van Caekenberghe <[hidden email]> wrote: > > > >> On 1 Jun 2020, at 06:39, Jerry Kott <[hidden email]> wrote: >> >> If you read the text of the EAR and take into account all other facts, I think that the notion that anyone should get into trouble by copying open source Smalltalk crypto libraries to other repositories is just a pure FUD. This software is open source, it is publicly available including the source code, it is hosted on a domain that is controlled by a non-US entity, and it’s easily accessible in its current form from countries that are currently on the US ‘vorboten’ list. > > Indeed. > > |
In reply to this post by Paul DeBruicker
Hi Paul,
thanks for the info but I won't read it. I try to focus on things that matter. There are too many things that try to distract everyone from producing something helpful. Law suits are IMHO not of that kind. I would rather put some money on the table for someone building proper FFI to openssl. The Cryptography was a mess I cleaned a bit but I had also scenarios where the squeak code was not working properly. And there is no real reason to have a smalltalk implementation of crypto if we carry around openssl anyway (for iceberg and secure connects). PierceNg has an implementation that implements a subset of openssl. This implementation is modeled after the library so lots of class methods. I'd prefer to have something more object model like. So if you think you can implement this please contact me and tell me what you think how long it takes and how much it will cost to do at least the things we have now in Cryptography. I'm willing to collect money or pay it myself. Norbert
|
On Tue, Jun 2, 2020 at 6:18 AM Norbert Hartl <[hidden email]> wrote:
I second this. VA Smalltalk included a complete FFI wrapper of OpenSSL in VAST 8.6.2 (2015) and it was one of the best decisions ever. Mariano Martinez Peck Email: [hidden email] Twitter: @MartinezPeck |
In reply to this post by Stéphane Ducasse
Stef,
that is a second problem. The main problem to me is solving the crypto stuff once and for all. And then I would like to know why people like Ron say things like this. But yes, shouldn't go without explanation/reaction! Norbert
|
In reply to this post by NorbertHartl
On Tue, Jun 02, 2020 at 11:17:30AM +0200, Norbert Hartl wrote:
> PierceNg has an implementation that implements a subset of openssl. > This implementation is modeled after the library so lots of class > methods. I'd prefer to have something more object model like. A very small subset currently, as my original need was to create an X509 request in code. PRs welcome. https://github.com/PierceNg/OpenSSL-Pharo Pierce |
> Am 04.06.2020 um 02:44 schrieb Pierce Ng <[hidden email]>: > > On Tue, Jun 02, 2020 at 11:17:30AM +0200, Norbert Hartl wrote: >> PierceNg has an implementation that implements a subset of openssl. >> This implementation is modeled after the library so lots of class >> methods. I'd prefer to have something more object model like. > > A very small subset currently, as my original need was to create an X509 > request in code. PRs welcome. > > https://github.com/PierceNg/OpenSSL-Pharo > Norbert |
On Thu, Jun 04, 2020 at 10:31:09AM +0200, Norbert Hartl wrote:
> > Am 04.06.2020 um 02:44 schrieb Pierce Ng <[hidden email]>: > > On Tue, Jun 02, 2020 at 11:17:30AM +0200, Norbert Hartl wrote: > >> PierceNg has an implementation that implements a subset of openssl. > >> This implementation is modeled after the library so lots of class > >> methods. I'd prefer to have something more object model like. > > > > A very small subset currently, as my original need was to create an X509 > > request in code. PRs welcome. > > > > https://github.com/PierceNg/OpenSSL-Pharo > > > If it is a small subset it might be feasible to talk about the approach taken. Sure. I don't see too many misplaced class methods myself. I've just loaded the package into a fresh image for a spin. My self-built VM uses OpenSSL 1.1.1 (instead of OpenSSL 1.0.x in the prebuilt VMs) and there are C API changes between those two OpenSSL versions that break many tests, basic things like XXX_create becoming XXX_new, XXX_init becoming XXX_reset etc. So I've created branches openssl_1_1 and openssl_1_0 that will match the versions used by Pharo. Pierce |
Free forum by Nabble | Edit this page |