Closed bdorney closed 5 years ago
As a temporary workaround, can you just generate the new pickle
file from the 3.2.2
tag, since it seems that tag was the last one to provide a working pickle file with the tools as they are?
Yes okay I’ll give that a shot. Thanks for suggestion.
To add some additional information on this.
@mexanick proposed the following via email:
please see the table below. These are first four lines of pickled files generated by old version of reg_interface (left) and new one (right). The only difference is in the line #3. So when you generate a new pickle file, just change the line #3 to the older (left) version.
<80>^Bccollections | <80>^Bccollections
OrderedDict | OrderedDict
q^A]q^B(]q^C(U^GGEM_AMCq^D(crw_reg | q^A]q^B(]q^C(U^GGEM_AMCq^D(creg_utils.reg_interface.common.reg_xml_parser
Node | Node
So I have done that both on eagle60
and eagle64
however this resolved one issue and then created what appears to be another shown in this elog entry.
The recipe has been updated to edit only the end of a line and not the whole line. Now it works.
The recipe has been updated to edit only the end of a line and not the whole line. Now it works.
Sorry I didn't understand. What changes need to be done now?
that's an old comment - if you remember, we iterated couple of times and reached an agreement on how one should modify the pickle
d file without introducing any show stopping characters in the file. Thus, I consider this issue as resolved and would like to close it.
Ah okay, just as a sanity check can you confirm for me that step 27 is still valid?
yes
Brief summary of issue
After we stripped
sca.py
,gbt.py
, andreg_interface.py
fromxhal
and migrated tools toreg_utils
and gem specific additions ontop of that inxhal
(post xhal tag 3.2.2 the CTP7 seems no longer able to understand the pickle file that is generated on the DAQ machine; additionallyreg_interface.py
on the CTP7 will not allow you to generate a pickle file on it.This is a critical problem since
gbt.py
is relying on thegem_amc_top.pickle
file to program the GBT's over the fiber link.Types of issue
Expected Behavior
GBT's
Current Behavior
Steps to Reproduce (for bugs)
reg_utils
andxhal
post packaging split.gbt.py 1 2 config <gbt config file for GBT2>
Example failure shown in this elog entry
Possible Solution (for bugs)
Unknown.
Context (for feature requests)
Presently rolled back the pickle file to what was last generated by xhal tag 3.2.2 but this only knows about 2 optohybrids in the address table and the CTP7 firmware now supports four (since v3.5.1).
This is pretty critical because with the superchamber construction the number of running OHv3c's that need gbt programming over the fiber link simultaneously are now more than 2 boards.
Your Environment
gem904qc8daq
zsh