I suppose next steps are three. (1) Contact John Meinel and see if he has advice (or wants to investigate himself). (2) Pursue the ValueErrors by hacking meliae to give me a better idea of what line failed (the "Traceback" line overwrites the information at the moment) and see if I can hack the dump to be loadable while still keeping value. (3) Investigate with Salgado if any of these sound familiar to him, and if he has a workaround.
Thanks to spm, we have seven memory dumps so far.
I have tried to analyze all of them with meliae. meliae failed to process all of them. The following are summaries of their failures.
launchpad- memory. 20566.dump: During loader.load, got "ValueError: Expecting object: line 1 column 17 (char 17)"
launchpad- memory. 20610.dump: During loader.load, got "ValueError: Unterminated string starting at: line 1 column 36 (char 36)"
launchpad- memory. 20654.dump: During om.compute_ parents( ), process died with no explanation ("Killed") and no core dump.
lpnet11.21094.dump: During om.collapse_ instance_ dicts() , process died with no explanation ("Killed") and no core dump.
lpnet12.21146.dump: ValueError: Expecting object: line 1 column 79 (char 79)
lpnet13.21203.dump: ValueError: Expecting , delimiter: line 1 column 186 (char 186)
lpnet14.21255.dump: ValueError: Expecting property name: line 1 column 64 (char 64)
I suppose next steps are three. (1) Contact John Meinel and see if he has advice (or wants to investigate himself). (2) Pursue the ValueErrors by hacking meliae to give me a better idea of what line failed (the "Traceback" line overwrites the information at the moment) and see if I can hack the dump to be loadable while still keeping value. (3) Investigate with Salgado if any of these sound familiar to him, and if he has a workaround.