Hi Christoph --

> what is about "Installer new merge: #ffi", should this work too? :-)

Of course not. You need some kind of version control here. The FFI that works with Squeak 5.3 can be different from the one that works in 6.0alpha. That's where Metacello helps.

Best,
Marcel
Am 04.10.2021 19:55:07 schrieb Thiede, Christoph <christoph.thiede at student.hpi.uni-potsdam.de>:
Hi Marcel,

what is about "Installer new merge: #ffi", should this work too? :-)

Best,
Christoph
Von: Squeak-dev <squeak-dev-bounces at lists.squeakfoundation.org> im Auftrag von Taeumel, Marcel
Gesendet: Dienstag, 28. September 2021 17:47:50
An: squeak-dev
Betreff: Re: [squeak-dev] Cannot load FFI on Squeak 6 alpha 20608/pi
 
Hi Tim --

Please use this to install FFI:

Metacello new
configuration: 'FFI';
load.

Unfortunately, ConfigurationOfFFI has outdated Metacello-load-code. "seaside.gemstone.com/ss/metacello" has not been working for ages. Yet, "http://seaside.gemtalksystems.com/ss/metacello" is still operational.

Best,
Marcel
Am 28.09.2021 17:42:26 schrieb Marcel Taeumel <marcel.taeumel at hpi.de>:
Hi all --

> [...] wants to fetch Gofer from http://seaside.gemstone.com/ss/metacello [http://seaside.gemstone.com/ss/metacello] which seems to be dead. Bit of a problem. [...]

Woah ... these days are quite challenging ... here is a Trunk image that smalltalkCI uses, which has Metacello pre-installed and can be updated without having to rely on the bootstrap code on that (dead) Squeaksource instance:

https://github.com/hpi-swa/smalltalkCI/releases/download/v2.9.6/Squeak64-trunk-20447.tar.gz


> Now so far as I can see it appears that the initial # in front of the <cdecl is no longer accepted? Is that the case?

I don't recall that it was ever supported. What should it mean? I checked the senders of #pragmaSequence in Squeak 5.2 and couldn't find any support for "#<cdecl:"

Best,
Marcel
Am 28.09.2021 16:27:13 schrieb Eliot Miranda <eliot.miranda at gmail.com>:

> On Sep 27, 2021, at 11:36 AM, tim Rowledge wrote:
>
> Trying to load FFI to check my old GPIO hardware driving code.
>
> It needs FFI. FFI has a ConfigurationOfxxxx. Load that.
>
> Run `ConfigurationOfFFI load`
> Nothing - because it is trying to do the #ensureMetacello thing and wants to fetch Gofer from http://seaside.gemstone.com/ss/metacello which seems to be dead. Bit of a problem.
>
> If I dig a bit I find that it seems we want to load ConfigurationOfFFI>>#version20:
> Loading the 'FFI-Pools-mt.31' , 'FFI-Kernel-mt.180' , 'FFI-Libraries-mt.2' & 'FFI-Callbacks-mt.25'; seems to work, then 'FFI-Tools-mt.37' is ok.
>
> Try to run the FFI tests and BOOM! So evidently sometihng isn't terribly happy.
>
> Going through all the install again to at least try my actual package results in a syntax error
> id: i i2cClose: handle
> "Disconnect from the i2c device, freeing the handle etc.
> returns 0 if ok, otherwise GpioBadHandle"
> #"'i2c_close' (long ulong) module: 'pigpiod_if2'>
> ^self externalCallFailed
>
> Now so far as I can see it appears that the initial # in front of the

I’m surprised it ever was. Is this some pre pragma syntax I’m ignorant of?

Indeed Tim, follow the syntax in the FFITestLibrary. Lots of examples there.

>
> tim
> --
> tim Rowledge; tim at rowledge.org; http://www.rowledge.org/tim
> A hacker does for love what others would not do for money.
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.squeakfoundation.org/pipermail/squeak-dev/attachments/20211005/61c3459f/attachment.html>