J. Kissel

Nine for the mortal men, doomed to die...

Another successful evening of CDS bug-finding. The h1sush34 computer has locked up, and the data concentrator / DAQ / framebuilder is dead because of my further attempts to make front end model changes to the SUS. 
YOU'RE WELCOME. 
Details and status below.

----
After completing the pending model changes on the HXTSs, I began recompiling, reinstalling, restarting, and restoring (4R'ing) all SUS on each of there respective front ends. Here's the rather gloomy status report:
             Main Library Part       Changes complete       Successfully Compiles        Installed         FE Process Restarted            Clean DAQ Status
h1suspr3        HLTS_MASTER               yes                       yes                     yes                    yes                            no
h1sussr3        HLTS_MASTER               yes                       yes                     yes                    yes                            yes
 
h1susmc1        HSTS_MASTER               yes                       yes                     yes                    yes                            no
h1susmc3        HSTS_MASTER               yes                       yes                     yes                    yes                            no
h1suspr2        HSTS_MASTER               yes                       yes                     yes                    yes                            no
h1sussr2        HSTS_MASTER               yes                       yes                     yes                    no(see detail 1)               no

h1susmc2        MC_MASTER                 yes                       no(see detail 2)        n/a                    n/a                            n/a
h1susprm        MC_MASTER                 yes                       no(see detail 2)        n/a                    n/a                            n/a
h1sussrm        MC_MASTER                 yes                       no(see detail 2)        n/a                    n/a                            n/a


Detail 1:
After successfully modifying and 4R'ing 3 of the 4 HSTSs which use the HSTS_MASTER, I figured the last would be a breeze. Of course not. Doing things no differently than on he other three, I successfully compiled and installed the h1sussr2 front end code. However, when I restarted the front-process,  the restart hung in the middle of turning on EPICs communication, and proceeded to crash the front end, and take down the entire data concentrator / DAQ / frame builder. I have a feeling that this is because of an error similar to what we found last Thursday on the h1sush56 computer (see LHO aLOG 8335). Note that h1sussr2 lives on a different computer -- h1sush34.

Detail 2:
Each of the models using the MC_MASTER library part complain the following same error snippet during compilation which is beyond me:

/opt/rtcds/lho/h1/rtbuild-2.7.2/src/fe/h1susmc2/h1susmc2.c: In function 'feCode':
/opt/rtcds/lho/h1/rtbuild-2.7.2/src/fe/h1susmc2/h1susmc2.c:1718: error: 'adc0' undeclared (first use in this function)
/opt/rtcds/lho/h1/rtbuild-2.7.2/src/fe/h1susmc2/h1susmc2.c:1718: error: (Each undeclared identifier is reported only once
/opt/rtcds/lho/h1/rtbuild-2.7.2/src/fe/h1susmc2/h1susmc2.c:1718: error: for each function it appears in.)
make[3]: *** [/opt/rtcds/lho/h1/rtbuild-2.7.2/src/fe/h1susmc2/h1susmc2.o] Error 1
make[2]: *** [_module_/opt/rtcds/lho/h1/rtbuild-2.7.2/src/fe/h1susmc2] Error 2
make[1]: *** [default] Error 2

This smells like it has something to do with the connections to the ADC0 parts, because there is certainly an ADC0 in every one of these models -- that the error says adc0 is undeclared must be a red herring. I've got an email out to the relevant CDS crew, I hope to hear from them and fix these problems by the morning.

I have still not begun MEDM screen modifications. I will do so once we resurrect each of the nine mortal men.