At 7:22 PM +0100 8/3/04, Alberto Manuel Brandao Simoes wrote:
Although I sent this ideas to this list, with another subject, here they go again to enter in the correct thread.

I think that to have ICU included both in CVS and tarballs created for parrot is not the best way. I know Dan (and maybe some others) do not want to depend on too much libraries. The main problem should not be on depend on ICU, but ICU not being usual on most systems.

So, we are scripting people (or this wouldn't be a perl mailing list). Then, we should be able to make Configure to test if ICU is in the system. If not, use the Internet connection to get it, untar, configure and compile. I think that nowadays it is very difficult to find someone trying to compile parrot without a network connection.

While a not unreasonable way to look at it, it's untenable. Counting on a network connection won't cut it, and part of the reason ICU's in CVS now is so that we can make sure things work properly. Obviously they aren't, so we should work this out while we can.


The right answer is to make ICU optional, and use the system ICU if it's available, with what we ship as a fallback as a last resort. I'm all for putting in patches for that, as well as making ICU properly optional again. I'd do that now but I'm tight for time for the next few weeks, so that's just not going to happen no matter how much I want it to.

If someone's got the time and I spec out the encoding and charset APIs, I'd be thrilled if ICU became optional again. Wouldn't hurt my feelings at all. We need it, because we need Unicode, but it doesn't have to be required.
--
Dan


--------------------------------------it's like this-------------------
Dan Sugalski                          even samurai
[EMAIL PROTECTED]                         have teddy bears and even
                                      teddy bears get drunk

Reply via email to