RAB on VAX and Alpha

[Antoniov.]

Vir,
I compared all values of RAB/FAB structures with my c code of indexed files. No any strange value I observed. Some field are pointer to other structure, so they numeric value are not important.
Just I understand your file has primary key of 4 bytes.
I looked at my old documentation and I confirm about 65537 status. This is a successfull code but I met it when rms routine could not get all information about file; this may happen when you open file throught network or something like that. I know at now it's not important.

Antonio Vigliotti

[Vir Thanvi]

I just tries the reverse compatibility and surprisingly it worked. I created a relative file on Alpha and copied it to VAX. I accessed the same file without any problem.

But if I create a file on VAX and copy it to Alpha and try to access it, I am still not being able to do that.

Antonio, I think you are trying with Indexed file but mine is relative. Second, my binary file had no record greater than INTEGER*4 or REAL*8. I will double check and repost a note, if I find anything else.

Thanks to all who are devoting precious time to help me out.

Best Regards,
Vir

[Volker Halle]

Vir,

if you create a file on Alpha and try to access it on Alpha, does it work ?

Volker.

[Robert Brooks_1]

Vir wrote . . .

Can someone please post a note about how can I obtain support from HP over this topic.



Your site will need to have a current software support agreement with HP. Do you know if your site has such an agreement?

-- Rob

[Vir Thanvi]

If I create a file on Alpha and access it on Alpha, it works fine.

I am not sure if my site has software support agreement with HP or not? I will inquire about that.

More info to add: I dumped the data on VAX and Alpha simultaneously and both seem to be reading identical blocks, but very strange thing happened. I was in loop to read data from the file with DO I=1,4. At each step I examined DATA(I). Initially when I examined DATA(2), it was all zeros (0). Then I examined DATA(3), it was filled up and then I re-examined DATA(2) and that had some values in it now. DATA(2) and DATA(3) looked the same on VAX and Alpha. Is it some timing issue, just wondering?

Best Regards,
Vir

[Antoniov.]

Vir,

I think you are trying with Indexed file but mine is relative. Second, my binary file had no record greater

I thought to indexed file because in your posted dump of Aug 15, 2005 17:11:47 GMT, I see RAB$B_KSZ=4; this is key size of RMS file.
I never used relative access, so I can't help you for this feature. I guess RAB$B_KSZ in relative access is size of pointer but I'm not sure about this.

Antonio Vigliotti

[Antoniov.]

RAB$B_KSZ is size of pointer in relative access.
See http://h71000.www7.hp.com/doc/731FINAL/4523/4523pro_010.html#99_rabb_kszfield
So you are reading on both platforms via 32 bit pointers.

Antonio Vigliotti

[Vir Thanvi]

Thanks to you all who helped in this process of extensive analysis and enhancing my system level knowledge. The problem was located at very top level where the tim-tag on file was being read incorrectly by a REAL*8 variable. As soon as I changed the REAL*8 to REAL*4, everything worked great.

Once again thanks for your time, effort and help.

Best Regards,
Vir

[Vir Thanvi]

Modified REAL*8 variables to REAL*4 as D-floating point double precision numbers are not supported on Alpha.