Begin forwarded message:

From: "David T. Lewis" <lewis@mail.msen.com> Subject: Re: [Pharo-project] OSProcess - is process done? Date: April 24, 2012 8:59:03 PM EDT To: Pharo-project@lists.gforge.inria.fr, sean@clipperadams.com

On Fri, Apr 20, 2012 at 03:02:38PM -0700, Sean P. DeNigris wrote:

...

Here's the OSProcess test results:

* Pharo 1.3 * Mac Lion 10.7.3 * recent Cog VM from the Pharo CI server * latest version of OSProcess (the all-in-one package, not the broken up ones)

192 run, 110 passes, 0 expected failures, 79 failures, 3 errors, 0 unexpected passes Is that what you would expect?!

Hi Sean,

I do not have a Mac, so I cannot really reproduce what you are doing. But for what it's worth, I did some testing on my Linux machine with Pharo 1.4.

OSProcess and CommandShell will run on Pharo 1.4 although there are a few problems to be sorted out. The attached screen shot shows results of a full test run using a 64-bit interpreter VM on Linux with the three required plugins (OSProcessPlugin, AioPlugin, and XDisplayControlPlugin).

I made some updates to both OSProcess and CommandShell to address issues flushed out by Pharo 1.4. These are minor updates, but you may want to update to the latest versions from SqueakSource.

Cog VM distributions do not include AioPlugin or XDisplayControlPlugin, and a portion of the OSProcess/CommandShell test suite will fail because of this. The AioPlugin should work with Cog but is not distributed. XDisplayControlPlugin does not yet work with Cog and it probably will never work on OS X. Most OSProcess functionality will work without these plugins, so the main impact is just on the test suite.

Pharo 1.4 deprecates #withBlanksTrimmed and #withoutTrailingBlanks, and suggests replacing them with Pharo methods that do not exist in other images. I do not know what to do about this for OSProcess/CommandShell, so I just commented out the deprecations in order to run the tests (turning off the deprecation warning preferences does not work, see below).

Something seems to be broken in the interaction of deprecation warnings and the test runner. If I do not comment out the deprecations in the actual methods, the test runner will show tests as failing (presumably because an exception was thrown), but they pass when I try to debug them (so maybe the exception is being handled normally when debugging, and is ignored?). As noted above, I just commented out the deprecations in the methods to make the problem go away.

HTH, Dave