<div dir="ltr"><br><br><div class="gmail_quote">On Thu, Oct 16, 2008 at 12:05 PM, Mark Stosberg <span dir="ltr"><<a href="mailto:mark@summersault.com">mark@summersault.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="Ih2E3d"><br>
> Mark: if you're reading this, I wanted to thank you for all the hard work<br>
> if you've put into setting up the Perl testing infrastructure. It had some<br>
> really cool features, but I think that our preference for simpler tests<br>
> and simpler infrastructure outweighs our desire for features. The recent<br>
> licensing issues and Windows buildbot issues were basically just piling on<br>
> that extra bit of motivation to go ahead and make this change. Anyway,<br>
> thanks again! I hope to see lots of nice shell tests from you (and hey<br>
> hopefully some darcs code?) in the future.<br>
<br>
</div>I appreciate that David let me write tests in Perl in the first place.<br>
<br>
I think that was a strategic move, because it allowed someone to contribute and<br>
help the project who otherwise wouldn't and increased the quality of the<br>
project at that time.</blockquote><div><br>Thanks Mark. I speculate that at that time no one foresaw the the reasons to switch that we have now. <br><br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
I also understand that new concerns are coming up now, and I appreciate the<br>
benefits of simplicity. I also respect that the current test suite maintainers<br>
may have different choices about the best tools to use, and I feel now like<br>
David may have felt at that the time, that the people actually doing the work<br>
should have some amount of freedom in how they implement the solution.</blockquote><div><br>Speaking of which...I think Zooko makes a really wise point when he recommends our tests be written in the same language we use, Haskell. The benefit is that we already require a Haskell compiler. This would certainly help address the test suite portability issues we have. But, I guess the big reasons to not do that are:<br>
1) It's a lot of work to convert all the tests to a new language.<br>2) It's really easy to tell our .sh scripts are doing in terms of real commands to reproduce it.<br><br>Then again, maybe the easier thing is to find a different shell for using on windows. Maybe there is a Haskell based bash-ism tolerant shell we could recommend for windows testers. <br>
<br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Overall, I'm glad the project is continuing to move forward, and I thank those<br>
who have helped in the past, and who continue to carry on the work.</blockquote><div><br>Yes. Progress is nice, and every time a test catches a bug I just want to thank each and every test contributor. It's really wonderful to have a test suite.<br>
<br>Thanks,<br>Jason<br></div></div><br></div>