Recover about cyber

[fubark]

Yes, you should be able to recover the stack trace for the implementation of recover. The current goal for libstd is to keep it small, and allow libs to be easily imported from the web... So one would do import json 'https://libs.com/json.cy'. But we could also keep an official set of modules in this repo. They would just be library modules that expose the interface with bindLib. There are a couple of things that needs to be done first though... one is user modules which I'm working on now. @matu3ba

[zachcp]

Awesome! Thank you.

[fubark]

Hello @zachcp, I just fixed coreReadLine in 0c59964 to return a error(#EndOfStream) value when EOF is encountered. recover hasn't been implemented yet as the design hasn't been finalized. This issue can remain open as reminder to complete recover. I also wouldn't recommend using try right now outside of writing tests. Here's how you might handle the error:

for:
    line = readLine()
    if line == error(#EndOfStream):
        print 'end of stream'
        break
    else:
        print line`

[matu3ba]

Will this also work for use cases like this:

def runTest(da: dalib.Talib) -> int:
    st = da.expectEq(list(ta.storage_msg[0].keys())[0], "1")
    # This line feels very superfluous, but I need stack trace (without crashing) + dump program state
    if st != 0: da.failTestDumpState(st); return 1 
    st = ta.expectEq(list(ta.storage_msg[0].keys())[1], "2")
    if st != 0: da.failTestDumpState(st); return 1 # again feeling superfluous (2x time code size)
    # ...
    def failTestDumpState(self: object, status: int):
        print("FAIL with status:", status)
        traceback.print_stack() # printStackTrace
        self.dumpState()
        print("last_setEvState:", self.timeline_msg[self.dump_timeline_msg_last])
        print("Messages still in Queue:")
        while self.queue_msg.empty() == False:
            msg= self.queue_msg.get(block=False)
            writeQueueMsg(msg)
da = dalib.Dalib()
da.init()
st = runTest(da)
sys.exit(st)

?

If @fubark provides sufficient pointers on how libstd works, then I can start with libstd xml and json/dictionary functionality I dearly miss in python or is super clunky. This should also provide better insights into performance of realistic applications.

[fubark]

recover is no longer part of the language since Cyber has moved to a try/catch mechanism. Please open a new issue if anyone has trouble using it. Learn more about error handling here: https://fubark.github.io/cyber/docs/toc/errors/