| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 1009.1 | | TLE::D_SMITH | | Wed Sep 13 1995 17:35 | 7 |
| 1009.2 | time frame for that ECO?? | CSC32::E_VAETH | | Thu Sep 14 1995 11:33 | 5 |
| 1009.3 | | TLE::D_SMITH | | Thu Sep 14 1995 11:56 | 5 |
| 1009.4 | Customer back... | CSC32::E_VAETH | | Fri Oct 27 1995 16:09 | 6 |
| 1009.5 | | TLE::D_SMITH | | Sat Oct 28 1995 09:32 | 3 |
| 1009.6 | Is there an OpenVMS Alpha V6.1 ECO for this problem? | SMAUG::GARROD | IBM Interconnect Engineering | Wed Mar 26 1997 19:24 | 65 |
| The same problem seems to exist on OpenVMS Alpha V6.1 and
also OpenVMS VAX V6.1.
What is the version of DECC$SHR.EXE that needs ro be run on V6.1
in order not to see this problem.
The image IO was running on OPenVMS Alpha V6.1 exhibits this problem
Image Identification Information
image name: "DECC$SHR"
image file identification: "T02.0-018"
image file build identification: "X5SC-SSB-0000"
link date/time: 13-APR-1994 16:08:48.76
linker identification: "T11-11"
as does the latest DECC ECO available. After installing this image
its ident is:
Image Identification Information
image name: "DECC$SHR"
image file identification: "T02.0-018"
image file build identification: "X5SC-SSB-BA00"
link date/time: 6-JAN-1997 17:35:36.67
linker identification: "T11-11"
The release notes for this ECO do not mention this fix for OpenVMS
Alpha.
In looking at all the release notes for current ECOs. It looks like
the following ECOs incorporate the fix:
OPENVMS_ALPACRT05_070_Alpha_V7_0_DEC_C_RTL_ECO_Summary
OpenVMS_ALPACRT08_062_Alpha_V6_2_-_V6_2-1H3_DEC_C_RTL_ECO_Summary
OpenVMS_VAXACRT05_070_VAX_V7_0_DEC_C_RTL_ECO_Summary
OpenVMS_VAXACRT08_062_VAX_V_6_2_DEC_C_RTL_ECO_Summary
OpenVMS_VAXACRT09_061_V5_5-2_through_V6_1_DEC_C_RTL_ECO_Summary
Note the distinct absence of a fix for OpenVMS Alpha V6.1. Specifically
ECO:
OpenVMS_ALPACRT09_061_Alpha_V6_1_-_V6_1-1H2_DEC_C_RTL_ECO_Summary
does not include the fix.
My questions are three:
1) Is there a supported DECC$SHR image for OpenVMS Alpha V6.1 that
includes this fix. If so how do customers get it.
2) Is OpenVMS Alpha V6.1 still formally supported by Digital? If
so I guess I can tell customers who come across this problem
in our producty that they HAVE to upgrade to V6.2 and then
apply the ECO. But I don't feel I can say this if V6.1 on Alpha
is still formally supported.
3) Why is there a fix for OpenVMS VAX V6.1 but not OpenVMS Alpha V6.1?
Regards,
Dave Garrod
IBM Interconnect Engineering
|
| 1009.7 | Program showing the problem | SMAUG::GARROD | IBM Interconnect Engineering | Wed Mar 26 1997 19:26 | 31 |
|
Attached is a noddy problem that demonstrates the problem.
foo.bar is any old file with characters in it. The prgram
fails when it attepts to unget() the 'B' character. It returns -1
instead of 'B'.
Dave
#include <stdio.h>
main ()
{
FILE *fd;
int status;
int i;
char c;
fd = fopen("foo.bar", "r");
for (i=0; i<3; i++)
{
c= getc(fd);
printf ("read character '%c'\n", c);
}
status = ungetc('A', fd);
printf ("status = %d\n", status);
status = ungetc('B', fd);
printf ("status = %d\n", status);
c = getc(fd);
printf ("read character '%c'\n", c);
return 0;
}
|
| 1009.8 | Maybe not the problem I'm seeing | SMAUG::GARROD | IBM Interconnect Engineering | Wed Mar 26 1997 20:40 | 13 |
| Further investigation shows that the ungetc() problem I'm having
doesn't seem to be fixed by ECO
VAXACRT09_061
on a V6.1 VAX. The ECO release notes say:
o The function ungetc stopped working in OpenVMS V6.1 for both
variable and fixed length records.
I'm going to open a new note on this (1283).
Dave
|
| 1009.9 | | SPECXN::DERAMO | Dan D'Eramo | Wed Mar 26 1997 20:43 | 22 |
| > Attached is a noddy problem that demonstrates the problem.
> foo.bar is any old file with characters in it. The prgram
> fails when it attepts to unget() the 'B' character. It returns -1
> instead of 'B'.
> status = ungetc('A', fd);
> printf ("status = %d\n", status);
> status = ungetc('B', fd);
> printf ("status = %d\n", status);
The standard says of ungetc
One character of pushback is guarenteed. If the
ungetc function is called too many times on the same
stream without an intervening read or file positioning
operation on that stream, the operation may fail.
Unless our documentation makes more of a committment than that
one character of pushback, then the behavior you are seeing is
legal for ungetc.
Dan
|