Comparing Rebol 3, Rebol 2, Rebol 1.3

Goodness knows, I am dealing with the fact that Rebol 2.7.6 doesn't run scripts written with 1.3.2....much less what Rebol 3 might do. I submitted a Rambo Report and was chastised for suggesting that a bug was in the pudding somewhere. My report was "dismissed". I notice that when I sent the entire script and its input file to the big guns on the scene.....so they could run it themselves....the chastisement ended....but of course, no resolution,  just "dismissal". John D.

On Thu, Mar 12, 2009 at 12:41 AM, John Dutcher <jfdutcher1958-yahoo.com> wrote:

> Goodness knows, I am dealing with the fact that Rebol 2.7.6 doesn't run scripts written with 1.3.2....much less what Rebol 3 might do. > > I submitted a Rambo Report and was chastised for suggesting that a bug was in the pudding somewhere. My report was "dismissed". I notice that when I sent the entire script and its input file to the big guns on the scene.....so they could run it themselves....the chastisement > ended....but of course, no resolution,=A0 just "dismissal".

On Thu, Mar 12, 2009 at 12:41 AM, John Dutcher <jfdutcher1958-yahoo.com> wrote:

> Goodness knows, I am dealing with the fact that Rebol 2.7.6 doesn't run scripts written with 1.3.2....much less what Rebol 3 might do.

The point of sending the file and script was not to get something fixed..... but to allow you to run both ways as I do, and see the dramatic difference in results...... a sort of 'proof' that the script has no flaw that needs fixing, only 2.7.6 has a flaw. John D.

GS> It's funny that all of us have tens of megabytes of code (I'm not GS> kidding, Qtask is like 17 MB) that had no problems at all. In John's defense, I think we've also had things break in other cases, and 2.7.6 *has* broken some of my stuff with the change to CALL. Other changes have also been problematic, e.g., "read %/" changing whether leading slashes are included, where prefs are loaded from (or if they are with unofficial releases), and registry functions being included or not. GS> Now, if you could point out to an example of what your problem is, GS> I guess it would be easy to solve. Sending in a complete script GS> and asking us to fix your code won't work, sorry. John posted his script and test file to the ML before posting to Rambo. It makes it easy to see that open/seek under 2.7.6 seems to also imply /binary, which previous releases did not. In this case, we're all right, and we're all to blame. John is right that REBOL broke something. He asked for help in narrowing it down, but we didn't, and he didn't on his own before posting to RAMBO. Ultimately, this isn't technically a bug, but it's a problem. The only place I found the behavior change mentioned was http://www.rebol.net/upnews/0026.html and, to find that, you have to know what you're looking for. The doc string for OPEN should mention it. I RAMBOed it. -- Gregg