[Info-vax] Porting logical name applications to Unix

[Alan Winston - SSRL Central Computing]

In article <00A92465.7C108184 at SendSpamHere.ORG>, VAXman-  @SendSpamHere.ORG writes:
>In article <00A923F6.23C7AB20 at SSRL.SLAC.STANFORD.EDU>, winston at SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing) writes:
>>In article <ceznAdBoufxH at eisner.encompasserve.org>, koehler at eisner.nospam.encompasserve.org (Bob Koehler) writes:
>>>In article <00A923BB.4839ED4B at SSRL.SLAC.STANFORD.EDU>, winston at SSRL.SLAC.STANFORD.EDU (Alan Winston - SSRL Central Computing) writes:
>>>> 
>>>> They *can*, but in fact (on VMS, even) whether your C program will immediately
>>>> start acting on the updated translation depends on the setting of a C 
>>>> runtime logical name.  (We got better performance from Apache when we had it
>>>> cache the translation rather than retranslating hundreds of thousands of
>>>> times).
>>>
>>>   What runtime logical name?  Unless the C RTL is going out of its way
>>>   to complicate things logical name translation for file opens are
>>>   done by RMS.
>>>
>>
>>DECC$ENABLE_GETENV_CACHE 
>>
>>
>>I wasn't talking about file name translation in particular, but logical names
>>in general.  But in the case of long-running C programs (probably the only case
>>you might want to enable a GETENV_CACHE, they'll typically open their files at
>>the beginning, which means you can change the logical name variable all you
>>want after the file is open in  your long-running program, but it's still going
>>to have the old file open.
>
>Do you set this system-wide (DEFINE/SYSTEM) or only for Apache (DEFINE/JOB)?

Only for Apache.  (Been a while since I messed with it, but I think I stuck 
the definition in the APACHE$WWW LOGIN.COM

-- Alan