[Pharo-project] Make things easy for newcomers!! (Was: Re: OSProcess and SqueakMap)

Ross Boylan ross at biostat.ucsf.edu
Mon Sep 7 22:30:47 EDT 2009


On Mon, 2009-09-07 at 17:50 -0700, Igor Stasenko wrote:
> A new blood in the herd! Wellcome.
Thank you.  I've used smalltalk and squeak for many years, albeit
intermittently, BTW.
> 
> As for the rest of gray-beards, i wonder, how long it to make a
> debugger window to pop-up with
> one additional button, named 'file a but report' so then any newcomers
> will send a proper stack trace instead
> of a screenshot and implementation of DNU method?
I thought the fileout option from the debugger would file out the stack.
There is also the mail bug report, but I was trying to keep in my main
email client (among other things, it would preserve the thread).

I'm just trying to get a working image with magma.  "Working" includes
being able to pass the magma test suite.  I thought to try Pharo after
running into trouble with 3.10.  It's frustrating that after 2 days I
don't have it working.  And I was thinking of trying Seaside with Magma
for something at work.

I think a newbie would have given up much earlier, though it's a little
hard to imagine a newbie wanting an object database...

This seems as good a place as any to report some other glitches, which
mostly strike me as documentation issues.  This is in addition to the
problem getting OSProcess going reported earlier in this thread, the
problem locating the sources file (which may be a buglet) and my VM
issues in other threads.

All of this is on Linux, and refers to
http://www.pharo-project.org/pharo-download.  I'm guessing life is
easier for Windows or Mac users.

First, it would be useful if that page clarified the VM requirements and
whether Pharo requires a fork of the VM (I take it one needs a VM with
block closures--which means pretty recent--but the VM itself is not
forked).  I also was unsure what would happen if I had different VMs
lying around, though that might be too esoteric to put on the page.

Oh, and there is no statement "you need image, changes, source, and a VM
and here is how they are related."

Second, the instructions say 
"When downloads are completed, uncompress all files in a folder (you
need an .image, .changes, and a .sources file) and run the virtual
machine."

I would interpret this as saying 
mkdir myproject
cd myproject
unzip /some/path/to/file.zip

But this actually creates a directory underneath my project that has
everything.

Then if you unzip the virtual machine, that ends up in a different
subdirectory.  Maybe the sources zip would go to a 3rd directory (I just
made a symlink to a copy I already had).

So then someone new to smalltalk would need guidance about how to put
these together and actually start the image.

Third, the image has an update facility which was not mentioned.  I
wasn't entirely clear whether updating was a good or bad idea.  I did
the update.  Some guidance on the install page would be helpful.

A final curiosity: I created symlinks so that the image and changes file
appeared directly under myproject, invoking the VM from that same
directory.  After I did the updates, the cache directory was under the
image subdirectory, i.e. as if I had run from the target of the symlink.

Ross









More information about the Pharo-dev mailing list