| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 956.1 | | TALLIS::wlfgng.amt.tay1.dec.com::nelson | It's not the years it's the mileage! | Fri Feb 28 1997 09:41 | 23 |
|
If you're going to report problems, you really need to also report what version
of VMS you're linking your objects on. Please verify that you're linking on OpenVMS
V5.4-2 (the only version of VMS that you can link your objects and translate them).
Be that as it may, I tried your example and discovered that it won't VEST;
please also report VEST errors that are reported. The following is a warning; we made
it that because it's possible this address might not actually get called. But if it
does, it could be your problem.
If I were debugging this, I would try to find out what message it trying to
get reported.
%VEST-W-UNSUPABSREF, Reference to unsupported absolute address (P1_space+000EDF
40)
At: CALLER\CALLER\16 [COBOL] (00000649)
Input: CALLG E4(R11),@#7FFEDF40
%VEST-W-TRANSWARN, Translation completed with warnings -- review them before usi
ng the output image
|
| 956.2 | Works for me | TALLIS::wlfgng.amt.tay1.dec.com::nelson | It's not the years it's the mileage! | Fri Feb 28 1997 16:12 | 6 |
| In addition, I just tried your sample and it worked just fine for me.
Brian
|
| 956.3 | VEST version which has problem | HANDVC::STEVELIU | | Mon Mar 03 1997 05:09 | 18 |
|
version of VEST I'm using is :
Image Identification Information
image name: "VEST_TV"
image file identification: "V1.1A-0"
link date/time: 13-OCT-1994 11:12:46.83
linker identification: "V1.1A-0"
Patch Information
There are no patches at this time.
which version are you using in .2 ?
|
| 956.4 | 1.1A | TALLIS::wlfgng.amt.tay1.dec.com::nelson | It's not the years it's the mileage! | Mon Mar 03 1997 11:47 | 12 |
|
I'm using V1.1A, which is the latest released version. It sounds like something isn't
set up correctly on your system; I can't think why else the same code would translate
fine for me and not for you.
A couple of folks have asked why you're not using the Alpha version of the COBOL
compiler; is there any technical reason not to use it?
Brian
|
| 956.5 | VESTing vs Recompilation | HANDVC::STEVELIU | | Tue Mar 04 1997 20:58 | 10 |
|
The reason we prefer to use VEST is because we have limited time
to migrate a large volume of Cobol applications to Alpha.
We tend to think VEST is the fastest way to get us there even if
Alpha Cobol compiler is available.
Any advices you could offer for us.
steve
|
| 956.6 | VAX Basic did some really strange things | HNDYMN::MCCARTHY | A Quinn Martin Production | Wed Mar 05 1997 05:57 | 7 |
| >> Any advices you could offer for us.
I believe that VESTing of any image containing BASIC code is likly to cause
problems and there are some areas of the language that simply will not
VEST correctly.
Brian J. - (who once worked on the port of BASRTL to DEC$BASRTL (Alpha))
|
| 956.7 | | TALLIS::wlfgng.amt.tay1.dec.com::nelson | It's not the years it's the mileage! | Wed Mar 05 1997 11:43 | 18 |
| It's been ages (first year in college) since I wrote COBOL, but my understanding is
that it's a VERY standard language. Certainly I would expect very few differences
between VAX COBOL and DEC COBOL (Alpha version).
That being the case, I would at LEAST investigate using the native compiler. If all
you have to do is recompile, you're golden. Plus you get native performance.
Brian (the other one) is right -- the VAX BASIC RTL has some problems, although I
think we weeded out quite a few. WHEN clauses are a case in point -- **READ** the
VEST release notes for what to do about WHEN clauses.
VEST was meant as a transition tool, not a long-term migration path. Now that all
the compilers have long since been ported, you really should be recompiling.
Brian
|
| 956.8 | Recompilation ?? | HANDVC::STEVELIU | | Wed Mar 05 1997 22:21 | 13 |
|
OK, then Is a VAX BASIC source file 100% compilable with DEC BASIC
compiler ? if not, what are the exceptions ? same question for
VAX and DEC Cobol ?
The original VAX sources are compiled by
-VAX BASIC V3.7 or earlier,
-VAX COBOL V5.3 or earlier,
which version of DEC BASIC and COBOL should I use to recompile ?
sl.
|
| 956.9 | ask in their own conferences | HNDYMN::MCCARTHY | A Quinn Martin Production | Thu Mar 06 1997 05:37 | 9 |
| >> OK, then Is a VAX BASIC source file 100% compilable with DEC BASIC
>> compiler ? if not, what are the exceptions ?
No its not. Check out TURRIS::DECBASIC for answers to those questions (it was
a while ago so I won't even try to put a sub-list here).
I think clt::cobol is the Cobol (VAX and DEC) versions.
bjm
|
| 956.10 | Vesting vs. Recompilation | HANDVC::STEVELIU | | Thu Mar 06 1997 23:03 | 23 |
|
With a immense volume of basic applications (>1,000,000 line of codes)
and so many differences found between VAX and DEC BASIC as listed in :
CLT::CLT$LIBRARY:[DECBASIC.KITS]BASIC012_RELEASE_NOTES.TXT;1
under section 4:
4 Differences Between DEC BASIC and VAX BASIC..........
4.1 VAX BASIC Features Not Planned For DEC BASIC......
4.2 Incompatible Features Between DEC BASIC and
VAX BASIC.........................................
4.3 DEC BASIC Features Not Available in VAX BASIC.....
can you tell me why recompilation still a better alternative than
using VEST to migrate the applications ? Some of these applications
were written more than 10 years ago.
I need to get every advices that I can get.
steve
|
| 956.11 | If VEST was an option, the native compiler would not be there | HNDYMN::MCCARTHY | A Quinn Martin Production | Fri Mar 07 1997 06:04 | 10 |
| >> can you tell me why recompilation still a better alternative than
>> using VEST to migrate the applications ?
Both Brian and me (the other Brian) have told you. There are potential
serious problems with VESTING VAX BASIC applications. Many of which will
not show up until run time. If you can be 100% sure that you never
use the things lised in the pages you describe, and never have an ON ERROR or
WHEN block anywhere then you MAY be safe.
Brian
|
| 956.12 | y | HANDVC::STEVELIU | | Sun Mar 09 1997 21:32 | 8 |
|
can you provide me a sample BASIC program with the use of ON ERROR and
WHEN clause that will illustrate the differences in run-time behaviour
? thanks.
steve
|
| 956.13 | its ALL possibly broken | HNDYMN::MCCARTHY | A Quinn Martin Production | Mon Mar 10 1997 06:09 | 14 |
| >> can you provide me a sample BASIC program with the use of ON ERROR and
>> WHEN clause that will illustrate the differences in run-time behaviour
>> ? thanks.
Its not a difference - its simply broken.
VAX Basic created call frames "on the fly", transfered control to them and
expected very specific behavior during an unwind (which is how "on error" and
when blocks were implemented on VAX). VEST simply couldn't keep track of
all the things BASRTL expected.
The entire run-time model was re-written for OpenVMS Alpha.
Brian J.
|
| 956.14 | On Error and When | HANDVC::STEVELIU | | Mon Mar 10 1997 21:09 | 16 |
|
Re: .13
so base on this fact, any BASIC source with the use of ON ERROR and
WHEN clause should be handled by recompilation, Vest cannot help in
this case, other than that VEST should have no problem, right ?
In the VEST V1.1A release notes section 3.14.2, it mentions when
the BASIC source has an WHEN ERROR construct, it must be translated
with the /OPTIMIZE=NOSCHEDULE qualifier. So I figure vesting is
not entirely impossible.
sl.
|