Re: RPMMakefile fails after starting Makefile.Release target: create_ProductVersionFile

What I fear is that maybe somehow I have set "PEGASUS_ROOT" to well. The
scripting in RPMMakefile is such that the definition of PEGASUS_ROOT is
actually changed during its execution. In particular, it appears to move
PEGASUS_ROOT from a path ending in "pegasus" to a path ending in
"tog-pegasus-2.9.1". Hence if for some reason the RPMMakefile really is
getting the value of PEGASUS_ROOT  from .bashrc , env_var.status, or
env_var_Linux.status after PEGASUS_ROOT has moved --- then it would seem
only rational for the build to fail in the fashion I mentioned.



On Wed, Mar 3, 2010 at 12:13 PM, Paul von Behren
<[email protected]>wrote:

> I have seen the messages John saw - I recall working around the problem,
> but don't recall the details.  I agree with him that this is not related to
> PEGASUS_PLATFORM.  Looking at mak/config.mak, I believe that an unexpected
> value for PEGASUS_ROOT will not cause any sort of error message, but will
> cause cause PLATFORM_FILES to be set incorrectly - which in turn cases
> VALID_PLATFORMS to have a value of ".".  This may be the first symtom of
> missetting PEGASUS_ROOT - so double check that you have that set to the base
> of the pegasus distribution and that the mak sub-directory is readable.
>
> I am now recalling this symptom showing up when I has pegasus source on a
> shared directory - the "~" shell operator was now working as expected and I
> have to specify the full path in PEGASUS_ROOT.  But this is a fuzzy
> recollection.
>
> Try running make with the "-d" (debug)  or "-n" (dry-run) option to see how
> make is interpreting the internal variables.
>
> Hope this helps in troubleshooting!  Just keep telling yourself "make is
> fun" :-)
>
> Paul
>
>
> On 3/3/2010 9:20 AM, John Robinson wrote:
>
> Sahana,
> PEGASUS_PLATFORM was set in both the .bashrc script, the env_var.status,
> and the env_var_Linux.status script. The env_var_Linux.status and the
> env_var.status are linked to the same file.
> Before compilation, a check was made by executing "env" to verify that PEGASUS_PLATFORM
> was set.
>
> The logic of the line and the printout of the line itself verifies that the
> VALID_PLATFORMS was not set. The error line I posted was not edited. The
> output from the echo of VALID_PLATFORMS  was a single period: "." .
>
>
>
> On Wed, Mar 3, 2010 at 12:42 AM, Sampige, Sahana Prabhakar (STSD) <
> [email protected]> wrote:
>
>>  Hi John,
>> Looks like the PEGASUS_PATFORM variable has not been set through your
>> .bashrc script. Please check by doing an
>> # echo $PEGASUS_PLATFORM.
>> You can get a list of VALID_PLATFORMS from
>> http://cvs.opengroup.org/cgi-bin/viewcvs.cgi/pegasus/doc/BuildAndReleaseOptions.html
>>
>> VALID_PLATFORMS not being displayed is not the reason for your build
>> failure. Just that the error message displayed is incorrect.
>>
>> Hope this helps.
>>
>> Rgds,
>> Sahana
>>
>>  ------------------------------
>> *From:* John Robinson [mailto:[email protected]]
>> *Sent:* Wednesday, March 03, 2010 3:48 AM
>> *To:* [email protected]
>> *Subject:* RPMMakefile fails after starting Makefile.Release target:
>> create_ProductVersionFile
>>
>>
>>
>> make[1]: Entering directory `/usr/src/redhat/BUILD/tog-pegasus-2.9.1'
>> /usr/src/redhat/BUILD/tog-pegasus-2.9.1/mak/config.mak:317: ***
>> PEGASUS_PLATFORM environment variable must be set to one of the following:
>> .  Stop.
>>
>> Since the PEGASUS_PLATFORM  is set in both the unix environment from a
>> .bashrc setting and in the env_var_Linux.status file, I am left with the
>> conclusion
>> that for some reason the variable VALID_PLATFORMS is not being set.
>>
>> Any help offered would be appreciated.
>>
>>
>
>