On Wed, 18 Apr 2001, Craig A. Berry wrote:
> At 12:02 PM 4/18/2001 -0700, Peter Prymmer wrote:
> >On Wed, 18 Apr 2001, Martin Vorlaender wrote:
> >
> >>
> >> Errr, wait a moment:
> >>
> >> $ write sys$output f$message(%X10B90003)
> >> %CC-I-NORMAL, normal successful completion
>
> Doh! Of course an informational doesn't halt the compiler. I suspect the
> success status we are seeing indicates "compiled with informational
> messages" or similar.
>
> >> $ IF compile_status .EQ. %X10B90003 THEN compile_status = good_compile !!!
> >> $ GOSUB just_mcr_it
> >> $ longdblsize = tmp
> >> $ d_longdbl = "define"
> >> $ echo "Your long doubles are ''longdblsize' bytes long."
> >> $ ENDIF
>
> Much better. Is there any successful compile status that we *don't* want to
> be considered a good compile? Or in other words why check for a specific
> status rather than success?
A good question. The current setup is restrictive in what it accepts,
which is generaly "a good thing(tm)". To open the floodgates we could go
so far as to just check for good $severity. Unfortunately linking against
shareable images such as perlshere.exe (or e.g. Sybase DLLS) that have
been built from components that issued at least warnings tends to re-issue
warnings on LINK. I am not sure about informationals though.
> Perhaps we need to modify it to something like the following:
>
> ># install ought not need a source, but it doesn't work if one's not
> ># there. Go figure...
> >install : $(MINIPERL_EXE)
> > @ @perl_setup.com
> > If F$TrnLnm("Sys") .nes. "" Then Deass SYS
> > create/directory perl_root:[lib]
>
> This breaks if perl_root:[000000] does not exist yet. I seem to remember we
> require its creation prior to installation in the docs, but I don't think
> it's actually necessary in the current incarnation of installperl, at least
> for the non-CLD case.
I wrote "create/directory perl_root:[lib]" as a short hand for going
through the whole thing:
perl_rootless = (F$TrnLnm("PERL_ROOT") - ".]") + "]"
create/directory 'perl_rootless
or - whatever - it'll need to be tested in any event.
> > copy $(DBG)perl$(E) perl_root:[000000]
> > copy $(DBG)perlshr$(E) perl_root:[000000]
> > $(MINIPERL) installperl
> >
> >And make suitable changes to installperl so as not to double install
> >perl.exe and perlshr.exe (I suspect it'd be nice to avoid PURGE).
>
> It generally checks dates as well as existence before moving files.
Great.
BTW I am trying to do away with passing the $(extensions) macro to
make_ext.com in the perl descrip.mms since it is now exceeding 252
characters -- details in a bit.
Since my bandwidth is low: I've downloaded your zip kit but have not
played with it yet. To accomodate MMS and MMK automatically I was
wondering if you might run the same tests configure.com does for MMS/MMK
or perhaps do an f$type(MMK) to see if it's available then issue you
commands in terms of:
$ build_cmd := MMS
$ if f$type(MMK) .nes. "" then build_cmd := MMK
$! ... etc.
or somesuch...
Peter Prymmer
