Added -q --sequential option for running the program single-threaded. #14
Conversation
This tool can be very memory intensive, enough so to cause the tool to fail to finish executing on less powerful hardware. This modification allows the user to trade slower operation for less memory consumption, which is necessary for some uses. In my case, this option takes about 3x longer to run, using about 10x less memory space. Additionally, this is a suitable workaround for issue JohannesEbke#7.
|
Your workaround for the managedblockchain error still leaves the tests failing: I think managedblockchain has to be added to the SERVICE_BLACKLIST and I have updated my pull request accordingly. #12 |
PR JohannesEbke#12 uses this solution and travis is happy, so I want a happy Travis too
yapf fails and complains about my linesplit, so here's the same thing as one line, like it asks for.
|
Passes tests now. To be clear, f0c7706 (and 6e04ba7 depending on what you're doing with yapf) is the sole commit containing a useful contribution, and I'll happily resubmit a PR rebased from any passing commit @JohannesEbke likes. I'd like to see this option pulled in |
Inspired and generalized from @bcknimbis PR #14.
|
Hi @bkcnimbis ; thanks for the Idea and all the work! I've picked up the idea, but implemented it as a "--parallel N" flag, where N is the number of parallel requests in flight. I've left the pool in, since #7 has been fixed otherwise. |
This is a nice tool but wasn't working for me, so I made it work for me. The option to run single-threaded is very useful where memory usage is a concern and running time isn't. Additionally, this addresses issue #7 by cutting out the call to multiprocessing methods when the option is specified.
The managedblockchain commit was necessary for the tool to run on my machine, so it is included here as well.