On Thu, 2006-08-03 at 13:16 -0500, Craig A. Berry wrote: > At 6:25 PM +0100 8/3/06, Martin J. Evans wrote: > > > >Using vmstar with -o did the trick. > > > >and DBI passes the test. > > > >Unfortunately mms clean produces: > > > >$ mms clean > > > >MCR annar$dka0:[sys0.syscommon.perl5_8_6]perl.exe "-MExtUtils::Command" -e > >rm_f > > *.olb perl.c core core.[0-9] [.blib.a > >rch.auto.DBI]extralibs.all core.[0-9][0-9] DBI.bso dbi.c > >%SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual > >address=0000000001010101, PC=000000007AE26F60, PS=0000001B > > > > That ain't good. Can you reproduce it with something simpler, like:
No, that is what I thought. In fact, I was starting to mistrust HP's binaries so I've started a perl build from source. > $ perl "-MExtUtils::Command" -e "rm_f" foo.[0-9][0-9] I'll try tomorrow when I'm back at work. Could do it from here but I had so much set up in my session to get this far I'll wait. > I'm speculating that a regular expression has made its way into a > filespec where it doesn't belong. Of course it shouldn't crash Perl > in any case. I can't reproduce the crash with Perl 5.8.4 nor with a > current development build of Perl on Alpha, which are the versions I > have handy at the moment. > > >Any idea how to delete files like those following which gnu tar created: > > > >dbi-1^.52-rc1.tar;1 > >perl-5^.8^.8.DIR;1 > >perl-5^.8.8;1 > > > >I cannot even set def into the dir one. > > > $ set process/parse=extended > $ delete dbi-1^.52-rc1.tar;1 Thanks - I was struggling with this. A long time ago (over 15 years ago) I seem to remember multinet creating funny filenames with dollars in them every time the case changed but you could at least type in the $. I am afraid it has been a long time since I did serious development work on VMS (and in any case it was mostly macro32 and Fortran) and I have not kept up (it was quite a surprise to see mixed case filenames). In fact on the rare occasions I log into VMS I have to sort of stop thinking and just type (hit keys) with what is obviously burnt into some long-term memory ;-) > etc. > > Somewhere in the doc set there used to be an OpenVMS Guide to > Extended File Specifications, but I think it's been assimilated now > into various and sundry other places. > > > > >(As you may have guessed by now, I'm new to VMS 8 on itanium although I used > >to > >use VMS 5.? on VAX and alpha a long time ago). > > > >Martin > >-- > >Martin J. Evans > >Easysoft Ltd, UK > >http://www.easysoft.com > > I hope this means your fine products are being ported to OpenVMS I64. <apologies if this gets too close to an advert - it is not intended to be> Nice of you to say so. The answer is yes and no. ODBC for RMS is already ported and available from our web site (BTW thanks to HP for a loan itanium machine and their very helpful people in the DSPP program). The actual job I'm looking at is getting an ODBC driver we did not write working behind DBD:::ODBC. I have few details at present and was just trying to get back up to speed. However, the experience we get from this exercise will undoubtedly lead to a release of the odbc-odbc bridge on OpenVMS Itanium. Talking of which: Our previous release of the odbc-odbc bridge (to access remote ODBC drivers from OpenVMS) linked directly to DBD::ODBC without an ODBC driver manager. I seem to remember you posted some changes to the perhaps poorly named unixODBC driver manager for OpenVMS which made it work on OpenVMS and I presume Nick included them (I'll check tomorrow). Could I just check with you whether you got unixODBC to build and work on OpenVMS? As always, thanks for the assistance - this is one very knowledgeable group which I've always found most helpful and I greatly appreciate the work of its contributors. I try to contribute as payback when I can but unfortunately I don't work on vms much these days. Martin -- Martin J. Evans Easysoft Limited http://www.easysoft.com
