Does it work to use Andy’s multipart version with encoding=”UTF-8” added to body elements?

On 8/14/17, 3:59 PM, "Giuseppe Celano" <celano@informatik.uni-leipzig.de> wrote:

    Thanks, Andy. I have also tried to invoke curl via proc:execute():

    proc:execute("curl",("-F", "data=@example.txt",  "-F", "tagger=", "-F", "parser=",  "https://urldefense.proofpoint.com/v2/url?u=http-3A__lindat.mff.cuni.cz_services_udpipe_api_process&d=DwIFAg&c=DS6PUFBBr_KiLo7Sjt3ljp5jaW5k2i9ijVXllEdOozc&r=JgwnBEpN1c-DDmq-Up2QMq9rrGyfWK0KtSpT7dxRglA&m=qRCvzBXhWnhavXBh6P8EXvNUf4YQi32_9ryqeOGjFo0&s=eR_QvfgwPhEJZno15HPXKZ7T6_aiGenJQ-NEFhJyCyU&e= " ))

    The function works, but unfortunately the text inside the file is not recognized as UTF-8, and so I get al lot of gibberish in the result. At the beginning I though it was due to
    my MacOS configuration, but I experimented a lot, and the problem seems to depend on BaseX.

    I run the basexgui (and basex) commands of the bin folder from my Terminal window and they should inherit the environment variables (and indeed proc:execute("locale") also shows the right UTF-8 values).

    I will open a Github issue, unless I am missing something here.