> On 08 Feb 2015, at 12:03, Manfred Kröhnert <[email protected]>
> wrote:
>
> Hello Sven,
>
> On Sun, Feb 8, 2015 at 11:32 AM, Sven Van Caekenberghe <[email protected]> wrote:
> It is always better to depend on Zinc #stable since it still evolves while
> Pharo 3 does not.
>
> thanks for your advice.
> This does mean that I should include the "spec project: 'Zinc-Http' with: [
> ..." lines, correct?
Yes, I think so, I have
project: 'ZincHTTPComponents' with: [
spec
className: 'ConfigurationOfZincHTTPComponents';
repository: 'http://mc.stfx.eu/ZincHTTPComponents';
loads: #('Core' 'Tests' 'REST' 'AWS') ];
which groups you load it up to you, adding the version is OK I guess (I think
it defaults to stable), I think the className is needed because of the odd
casing.
But I can't say I am a Metacello expert ...
> Thanks,
> Manfred
>
>
> > On 08 Feb 2015, at 11:08, Manfred Kröhnert <[email protected]>
> > wrote:
> >
> > Hello Dale,
> >
> > On Mon, Feb 2, 2015 at 10:52 PM, Dale Henrichs
> > <[email protected]> wrote:
> > Manfred,
> >
> > When you use github, you don't need nor do you want to use a
> > ConfigurationOf. git manages the versions for you so you don't need to
> > define versions in the ConfigurationOf. If you want versions for a git
> > repository, use tags.
> >
> > You still need to specify load order and project dependencies. For an
> > example you can look at[1] or download the baseline:
> >
> > Metacello new
> > baseline: 'External';
> > repository: 'github://dalehenrich/external:master/repository';
> > get.
> >
> > No need to specify a repository in the baselineOf, since the repository is
> > derived from the location of the Baseline itself.
> >
> > Dale
> >
> > [1]https://github.com/dalehenrich/external/blob/master/repository/BaselineOfExternal.package/BaselineOfExternal.class/instance/baseline..st
> >
> >
> > I have question regarding BaselineOf.
> > If my package is depending on the Zinc-HTTP package, would the following
> > baseline work for Pharo3 (which already includes Zinc)?
> >
> > baseline: spec
> > <baseline>
> > spec
> > package: 'HttpExplorer' with: [spec requires: 'Zinc-Http'];
> > yourself.
> >
> > Or is something like the following required in addition to make the
> > configuration complete:
> >
> > spec project: 'Zinc-Http' with: [
> > spec
> > repository: 'http://mc.stfx.eu/ZincHTTPComponents';
> > loads: #('Core');
> > version: #stable
> > ].
> >
> > Thanks,
> > Manfred
> >
> >
> > On 02/02/2015 01:21 PM, Manfred Kröhnert wrote:
> >> Hi,
> >>
> >> sorry for the unfinished email upfront. I accidentially hit the send
> >> button before I was finished...
> >> So here we go again:
> >>
> >> Some time ago I started a small Pharo based project but didn't want to
> >> announce it before I knew how to load it conveniently.
> >>
> >> For versioning I started using GitFileTree and pushed the code to GitHub
> >> together with a ConfigurationOf created with Versionner.
> >>
> >> Only after a more recent email from Kilon on this list I figured out that
> >> the project should be easily loadable with the following commands:
> >>
> >> Metacello new
> >> configuration:'HttpExplorer';
> >> repository:'github://mkroehnert/httpexplorer:master';
> >> get.
> >>
> >> Metacello new
> >> configuration:'HttpExplorer';
> >> repository:'github://mkroehnert/httpexplorer:master';
> >> load.
> >>
> >> Executing the first command succeeds and the ConfigurationOfHttpExplorer
> >> appears in the Browser.
> >> But when I try to execute the load command I get the following error
> >> (unfortunately I don't yet know how to copy a complete stacktrace):
> >>
> >> Error: Instances of UndefinedObject are not indexable
> >>
> >> This is a result from executing ZnClient>>getConnectionAndExecute where
> >> 'request url' returns the following address:
> >>
> >> http://:80/[email protected]:mkroehnert/httpexplorer.git/?C=M;O%3DD
> >>
> >> I guess that the error is probably in the ConfigurationOf but I am a bit
> >> lost as to where I should dig further.
> >> It would be great if anyone could point me in the right direction of where
> >> to look or tell what went wrong.
> >>
> >> Thanks in advance,
> >> Manfred
