Front page | perl.vmsperl |
Postings from June 2002
At 12:25 PM 6/12/2002 -0400, Henry A. Frystak - System Manager wrote: >><http://www.iki.fi/jhi/perl@17165.tgz> > Downloaded the above snapshot and >compiled it on my aging Vax system. Details are: VAX/VMS V5.5-2 on a VAXstation >4000-90a, Multinet V4.2 Rev A-X, DEC C V5.5-002 (yes, I know its old, but there >are reasons for not upgrading right now) and MMS V3.1-03. The @Configure went >O.K., but the following came up during the MMS : > >CC/DECC/Include=[]/Standard=Relaxed_ANSI/Prefix=All/Obj=VMS.obj/NoList/Define=PERL_CORE >VMS.C > STRLEN trnlen = strlen(trndir); > .............^ >%CC-E-NOSEMI, Missing ";". > At line number 3357 in $41$DKA400:[SYSSOFT.PERL.PERL5_8_0_RC1]VMS.C;1. Hmm. It looks like it doesn't want to create an automatic variable on the fly. It shouldn't hurt anything to move the declaration to the beginning of the function. Odd, since I've had no problem with DEC C 5.2 on Alpha. > _ckvmssts(sys$create_user_profile(&usrdsc,&usrprolst,0,0, > ..^ >%CC-I-IMPLICITFUNC, In this statement, the identifier >"sys$create_user_profile" is implicitly declared as a function. > At line number 6647 in $41$DKA400:[SYSSOFT.PERL.PERL5_8_0_RC1]VMS.C;1. Urk. I'm afraid this may be harder to fix. I found a VMS v5.0-era system services manual that does not include this function, nor does the description of sys$check_access have the user profile argument. Does anyone know when this support became available? The following command might tell us something, though I don't know if the image would be named the same on VAX or older VMS versions: $ search sys$library:secureshrp.exe sys$create_user_profile If the support is latent but undocumented in VMS v5.x, then we might be able to just supply a prototype, but if the support isn't there we'll either have to conditionally work around it or give up on supporting VMS versions that are 10 years old :-(. Oops, wait. The document here says it came along in v6.0: <http://www.compaq.com/support/asktima/appl_tools/00918BEC-4A9B9CA0-1C0069.html> So we'll need to go back to the simpler form of sys$check_access for older systems. >I will leave it up to the Perl and C experts to figure out what's wrong. With >the remaining fractions of a tuit left, I will try the snapshot on 2 other systems I >have access to, both Alphas of different types and flavors of o/s and software. Thanks, and thanks for the report. I'll try to work up a patch for the problems you've found but it's a little difficult making changes for systems I can't test.