Ben Coman wrote:


On 26 November 2017 at 21:28, Herby Vojčík <he...@mailbox.sk
<mailto:he...@mailbox.sk>> wrote:

    Ben Coman wrote:



        On 25 November 2017 at 15:18, Викентий Потапов
        <vikenti.pota...@gmail.com <mailto:vikenti.pota...@gmail.com>
        <mailto:vikenti.pota...@gmail.com
        <mailto:vikenti.pota...@gmail.com>>> wrote:


             I downloaded the latest PharoLauncher image, use VM for
        Pharo 6.1.
             1) On load i got an error with UTF8 encoding.
             2) Every time i tried to create image (for example, Pharo 7
        image) i
             get an error "Can't find the requested origin".

             Reported.

             Best regards, Vikenti Potapov


        The overall difficulty here in getting this addressed is
        difficulty in
        reproducing your environment.
        For example, I think it would be awkward for me to operate with OS
        locale set to Russian.
        You'll see little action on it until there is an Issue logged at
        pharo.fogbugz.com <http://pharo.fogbugz.com>
        <http://pharo.fogbugz.com/> with a simple set of steps
        to reproduced for standard environments (i.e. using English locales)

        Reviewing your other posts, I'd suggest initially creating a
        script to
        wrap Pharo startup that cleans all the environment variables to
        ascii


    -1.

    It should all work in other locales. Maybe you should set some if
    not all of the jenkins jobs under different locales (at least
    russian and japanese) to catch those things fast.

    Your (I hate to use the word) privilege to be fine with ASCII only
    actually masks lot of bugs from you.


I think you misunderstand me.  This suggestion was only to help isolate
the problem during troubleshooting.
The best way to troubleshoot something is to a.) start with something
working and then b.) be able to reliability break it.
Vikenti doesn't have (a.) so I was giving him a hack to get that, to be
able to proceed to (b.)

But your suggestion to the running some CI jobs under different locals
and timezones could be useful.

Ok, sorry.

Herby

cheers -ben



    Herby

    P.S.: Other masking thing is the timezone which is UTC in case of
    UK, France and lot of the rest of Western Europe. That should also
    be set to. say, Sydney and Honolulu to get big GMT+ and GMT- shifts.


        characters, temporarily creating required directories so at lest
        you can
        get it started.  Then extend that script to reintroduce the
        error, like
        a single non-ascii character in a single environment variable that
        causes an error.  That extended part then provides a path for
        others to
        reproduce your problem, which you would attach to the Issue.
          Then as
        you investigate a solution yourself, ask questions about it on
        pharo-dev
        to keep the Issue active and over time you'll likely get some action
        from others on it.

        cheers -ben








Reply via email to