Jordan, I hope you don't mind my cc'ing vmsperl; I think this is something that should be out there for general discussion.
At 4:23 PM -0400 4/22/02, Henderson, Jordan wrote: >I'm kind of the opinion that we should seriously think about formulating the >support we need from Compaq to do this stuff right and easily. The piping code >is ingenious, but I'm afraid it will be difficult to support. > >I think that PIPE driver, that's included but not support but is there in >OpenVMS, that's been mentioned on comp.os.vms by a Compaq Engineering guy (I >can't get to groups.google.com anymore), might be a good solution. OTOH, it's >only supported on reasonably new OpenVMS revisions. I would certainly welcome more involvement from OVMS engineering than the once-every-year-or-two post from Hoff. It's hard to think of a PowerPoint presentation they've made public in the last year or so that didn't have Perl as a bullet point in their ongoing plans, yet as far as I know they haven't done anything other than create the PCSI kit (and perhaps supported the mod_perl port?). We need, among other things, a complete review for ODS-5 and /PARSE_STYLE=EXTENDED and whatever that new parse style is that's coming in 7.3-1. I think it was Forrest Kenney who bragged in c.o.v. about cooking up the pipe driver in 20 hours over a Thanksgiving weekend, though he also said (but did not explain) that it was not appropriate for pipes in the C RTL. I have a feeling the official Compaq line in regard to our piping problems would be that they already provide popen() and pclose() in the C RTL and if something hangs then its our code, not theirs. Obviously a big customer who could demonstrate a business reason for giving the C RTL's pipes some attention and could show them exactly what doesn't work might get some action. That wouldn't be me. One thing we probably can do is make Chuck's piping implementation a configuration option; as it is the problems in the C RTL versions could get fixed and we'd never know it. -- ________________________________________ Craig A. Berry mailto:[EMAIL PROTECTED] "... getting out of a sonnet is much more difficult than getting in." Brad Leithauser
