Posted by
David T. Lewis on
Jun 28, 2013; 1:39am
URL: https://forum.world.st/Syntax-error-loading-OSProcessV4-3-7-sar-in-Cuis-4-1-1739-tp4695828p4695849.html
On Thu, Jun 27, 2013 at 04:19:05PM -0700, Casey Ransberger wrote:
> Aha. The wiki is out of date then and what else is new :)
>
>
http://wiki.squeak.org/squeak/708>
Way back when, I used to use the wiki to distribute the latest and greatest
OSProcess and CommandShell. Then someone came along and invented SqueakMap
and SqueakSource, so eventually I caved in to the pressure of moving to modern
technology ;-)
But to be serious, I'm quite sure that I updated that wiki page at squeak/708
(which is still the home page for OSProcess) several years ago to refer to
SqueakSource as the location of the current development. It does not say
that now, so I wonder if some wiki updates got lost somehow ... maybe someone
restored the wiki from backups and we missed a few things? Or maybe my memory
is failing me ... hmmm.
> So I went to squeaksource.com and grabbed the latest OSProcess and
> CommandShell. OSProcess reports 4.5.6, and CommandShell reports 4.6.5 (note
> the juxtaposition, not sure if they should report the same number or not.)
>
> Is this the usual way to install the current stable OSProcess into a Squeak
> image? Or should I use SqueakMap instead?
>
Either way is fine. SqueakSource is the master repository, so if you want to
be certain of getting the latest version, that's a safe bet until such time
as SqueakSource goes belly up. And assuming that I remember to keep SqueakMap
updated, that will do the same thing.
Dave
>
>
> On Thu, Jun 27, 2013 at 3:36 PM, David T. Lewis <
[hidden email]> wrote:
>
> > On Thu, Jun 27, 2013 at 03:18:07PM -0700, Casey Ransberger wrote:
> > > So after shuffling around the Monticello dance by changing .sar and .mcz
> > to
> > > .zip, nuking various files that aren't useful w/o MC, I'm having some
> > > trouble filing in the OSProcess source.st.
> > >
> > > I think I know what's going on here. Squeak probably still has
> > > BlockContexts for back compat, and Cuis 4.1 probably only has
> > BlockClosures
> > > now.
> > >
> > > In UnixProcess>>argsAsFlatArrayAndOffsets: I have this problem...
> > >
> > > (anArrayOfNullTerminatedStrings collect: [:e | e size])
> > > inject: offset into: [:p :e | ws nextPut: p. *Cannot store into ->*p :=
> > p +
> > > e].
> > >
> > > I don't think I've ever used #inject:into:, but looking at it, *I think*
> > > the above means the same thing as:
> > >
> > > (anArrayOfNullTerminatedStrings collect: [:e | e size])
> > > inject: offset into: [:p :e | ws nextPut: p. p + e].
> > >
> > > ...which compiles just fine. Am I right, or is that going to mean
> > something
> > > totally different? I'm not as familiar with the inject idiom, and I don't
> > > want to introduce a logic error.
> > >
> > > Thinking I want to use OSProcess to talk to the command-line git client
> > so
> > > I can make a nice little UI for it in Morphic.
> > >
> >
> > That's a bug in OSProcess, but it has probably already been fixed. The
> > current
> > version is:
> >
> > OSProcess versionString ==> '4.5.6'
> >
> > Dave
> >
> >
>
>
> --
> Casey Ransberger
_______________________________________________
Cuis mailing list
[hidden email]
http://jvuletich.org/mailman/listinfo/cuis_jvuletich.org