• Re: Configure country

    From druck@news@druck.org.uk to comp.sys.acorn.misc on Sat Oct 4 20:40:56 2025
    From Newsgroup: comp.sys.acorn.misc

    On 04/10/2025 18:47, Paul Sprangers wrote:
    In article <29189e665c.harriet@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    Presumably there are reasons why it would be inadvisable to leave
    *Alphabet permanently set to UTF8?

    It upsets some characters in menus, noticeably those that represent key
    short cuts.

    That's because those characters aren't part of code page and are only
    present in a few Acorn sourced fonts.

    ---druck
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Harriet Bazley@harriet@bazleyfamily.co.uk to comp.sys.acorn.misc on Sat Oct 4 20:35:31 2025
    From Newsgroup: comp.sys.acorn.misc

    On 4 Oct 2025 as I do recall,
    Paul Sprangers wrote:

    In article <410ea2665c.harriet@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    [...] didn't get Dict working with Russian
    input. Originally the issue was the lack of an appropriate dictionary server (the ones supplied in the download are no longer valid),

    Dict servers tend to come and go, which is a shame.
    Currently, these two Russian dicts seem to work:

    server pkgs4unix.org
    server dict.mova.org

    *experiments*

    Apparently in order to use Cyrillic in !Dict, as opposed to Netsurf, I
    need to set Alphabet to Latin1 and set Keyboard Russia (as opposed to
    UTF8 and UK)....

    Having successfully typed in a string to the reduced Dict window and
    pressed Return, I get a window popping up (with a jumbled Latin-1 title, presumably because, as I've noticed with Netsurf, RISC OS doesn't
    support non-Latin characters in title bars) which says

    1 definition found at pkgs4unix.org

    From -i-+-+-#-#-C-i -U-+-+-+-+-+-+-+-# -Y.-E-#-C-#-+-+-#-# [abr1w]:

    error: The request is not a valid UTF-8 string


    Same result from dict.mova.org.
    --
    Harriet Bazley == Loyaulte me lie ==

    A thing is not necessarily true because a man dies for it.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Harriet Bazley@harriet@bazleyfamily.co.uk to comp.sys.acorn.misc on Sat Oct 4 21:12:01 2025
    From Newsgroup: comp.sys.acorn.misc

    On 4 Oct 2025 as I do recall,
    Harriet Bazley wrote:

    [snip]

    Having successfully typed in a string to the reduced Dict window and
    pressed Return, I get a window popping up (with a jumbled Latin-1 title, presumably because, as I've noticed with Netsurf, RISC OS doesn't
    support non-Latin characters in title bars) which says

    1 definition found at pkgs4unix.org

    From -i-+-+-#-#-C-i -U-+-+-+-+-+-+-+-# -Y.-E-#-C-#-+-+-#-# [abr1w]:

    error: The request is not a valid UTF-8 string

    Ah, I think that's because at some point in the past I had configured
    !Dict to use Bukinist in its input icon. So while it displays Cyrillic
    at the RISC OS end, it isn't sending Unicode (apparently).

    If I reconfigure !Dict to use Cyberbit and switch keyboard mapping to
    Russ Uni, then for the string 'shut' (in Cyrillic) I get a variety of definitions, many of which are either *in* Russian or into languages I
    don't speak (Polish, Hebrew), but the third one down is 'jester, fool',
    which is the correct answer. :-)


    (I eventually managed to decipher the original word in the video that
    was giving me trouble by means of the traditional method of
    back-translation, i.e. guessing what it might have been from context and
    then looking up the possibilities in the English-Russian dictionary to
    see which words matched. Not 'khust', as I had heard it, but 'kholst',
    a painter's canvas - and no wonder I didn't recognise that piece of vocabulary!)
    --
    Harriet Bazley == Loyaulte me lie ==

    Eschew Obfuscation.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Paul Sprangers@Paul@sprie.nl to comp.sys.acorn.misc on Sat Oct 4 23:30:24 2025
    From Newsgroup: comp.sys.acorn.misc

    In article <89e7ab665c.harriet@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    Apparently in order to use Cyrillic in !Dict, as opposed to Netsurf, I
    need to set Alphabet to Latin1 and set Keyboard Russia (as opposed to
    UTF8 and UK)....

    That shouldn't be necessary at all. KeyMap produces cyrillic characters independent from alphabet or keyboard settings.

    Having successfully typed in a string to the reduced Dict window and
    pressed Return, I get a window popping up (with a jumbled Latin-1 title, presumably because, as I've noticed with Netsurf, RISC OS doesn't
    support non-Latin characters in title bars) which says

    1 definition found at pkgs4unix.org

    From -#]--a% b+[][+#]- +.#-a-#]-- [abr1w]:

    Have you configured a unicode font in Choices, under Bodytext? If so, it
    should show the suffix \EUTF8 automatically. It won't do so with Bukinist.

    Paul
    --
    https://riscos.sprie.nl
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Paul Sprangers@Paul@sprie.nl to comp.sys.acorn.misc on Sat Oct 4 23:33:20 2025
    From Newsgroup: comp.sys.acorn.misc

    In article <073faf665c.harriet@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    Ah, I think that's because at some point in the past I had configured
    !Dict to use Bukinist in its input icon.

    Ah, cross postings.
    But I will have a look at why F12 doesn't work when the virtual keyboard of KeyMap is open. I never noticed that.

    Paul
    --
    https://riscos.sprie.nl
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Harriet Bazley@harriet@bazleyfamily.co.uk to comp.sys.acorn.misc on Sun Oct 5 10:30:58 2025
    From Newsgroup: comp.sys.acorn.misc

    On 4 Oct 2025 as I do recall,
    Paul Sprangers wrote:

    I will have a look at why F12 doesn't work when the virtual keyboard of KeyMap is open. I never noticed that.

    That's quite a common problem in apps that intercept keypresses - so far
    as I recall you need to explicitly check for F12 and pass it on. I'm
    not sure I've written any programs with windows that do react to
    keypresses....

    Ah, here we are: SideDiff.

    If responding to a KeyPressed event in Wimp_Poll, the application needs
    to call Wimp_ProcessKey afterwards with any keypresses in which it is
    *not* interested in order to re-broadcast them around the desktop. Since
    KeyMap is designed to potentially map all keys on the keyboard to some
    other key (which might I suppose include the F12 key, if the user wished
    it!) it probably isn't making any exceptions.
    --
    Harriet Bazley == Loyaulte me lie ==

    Justice: A decision in your favour.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Harriet Bazley@harriet@bazleyfamily.co.uk to comp.sys.acorn.misc on Sun Oct 5 10:19:31 2025
    From Newsgroup: comp.sys.acorn.misc

    On 4 Oct 2025 as I do recall,
    Paul Sprangers wrote:

    In article <89e7ab665c.harriet@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    Apparently in order to use Cyrillic in !Dict, as opposed to Netsurf, I
    need to set Alphabet to Latin1 and set Keyboard Russia (as opposed to
    UTF8 and UK)....

    I think that was because in the distant past I had configured !Dict to
    to Bukinist as its icon input font (that being the only one I had
    managed to establish a working key-mapping for). With Cyberbit
    configured as input font I seem to be able to type into !Dict just by
    switching to RussUni

    (sigh: except that having done so, I'm still getting 'The request is
    not a valid UTF-8 string". I wonder if the issue is my attempt to use
    the arrow keys to edit a typo, which definitely seems to require
    multiple keypresses to move past each character? Or just that my
    Russian spelling is dodgy in the first place?)

    Edit: definitely using the arrow keys to try to delete things. I can reproduce that.


    That shouldn't be necessary at all. KeyMap produces cyrillic characters independent from alphabet or keyboard settings.

    Indeed, having configured StrongED to use Cyberbit as its display text
    for News mode while experimenting, I just accidentally typed a string of Cyrillic into this text without noticing! A feat I unfortunately seem currently able to reproduce, though I did notice the arrow key/deletion
    issues there as well... which makes sense, since the concept of 'one
    character per keypress' is probably hardwired all over RISC OS.

    I'm not sure using it in StrongED is a good strategy, as for some reason
    I seem to be producing far more typos than normal, though I don't quite
    see how the two can be connected! (I'm not used to using an outline
    font/bar cursor in StrongED at all.)

    Having successfully typed in a string to the reduced Dict window and pressed Return, I get a window popping up (with a jumbled Latin-1 title, presumably because, as I've noticed with Netsurf, RISC OS doesn't
    support non-Latin characters in title bars) which says

    1 definition found at pkgs4unix.org

    From |U|c|R|A|E|a|4 |i|y|Y|R|Y|y|L|R|A -+.-#|a|a|E|L|R|A|E [abr1w]:

    Have you configured a unicode font in Choices, under Bodytext? If so, it should show the suffix \EUTF8 automatically. It won't do so with Bukinist.

    Bodytext is configured to Cyberbit\EUTF8.

    But Bold and Italic are still set to Homerton\UTF8, which may be
    the problem here...

    Ah, that's better - I tried using DejaVu for both, and now get my Froms
    in valid Cyrillic. :-)
    --
    Harriet Bazley == Loyaulte me lie ==

    The nice thing about standards is that there are so many to choose from
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Paul Sprangers@Paul@sprie.nl to comp.sys.acorn.misc on Sun Oct 5 12:40:40 2025
    From Newsgroup: comp.sys.acorn.misc

    In article <4a64f8665c.harriet@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    That's quite a common problem in apps that intercept keypresses - so far
    as I recall you need to explicitly check for F12 and pass it on.

    Okay, KeyMap didn't react on keypresses at all, but now it does, just to
    pass on F12. Version 1.05 is now ready for download at www.riscos.sprie.nl

    Paul
    --
    https://riscos.sprie.nl
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Harriet Bazley@harriet@bazleyfamily.co.uk to comp.sys.acorn.misc on Sun Oct 5 12:46:15 2025
    From Newsgroup: comp.sys.acorn.misc

    On 5 Oct 2025 as I do recall,
    Paul Sprangers wrote:

    In article <4a64f8665c.harriet@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    That's quite a common problem in apps that intercept keypresses - so far
    as I recall you need to explicitly check for F12 and pass it on.

    Okay, KeyMap didn't react on keypresses at all, but now it does, just to
    pass on F12. Version 1.05 is now ready for download at www.riscos.sprie.nl

    Nice!

    *rereading manual for changes*
    I had forgotten about the ability to Toggle Size on the Keyboard window
    to display extra keys.

    (The download seems to have acquired a lot of 'odd' extra keymaps, e.g.
    'Test', 'Blank' and 'Woodtype', that you possibly didn't intend to
    include in a distribution version...?)
    --
    Harriet Bazley == Loyaulte me lie ==

    A man's best friend is his dogma.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Harriet Bazley@harriet@bazleyfamily.co.uk to comp.sys.acorn.misc on Sun Oct 5 13:27:20 2025
    From Newsgroup: comp.sys.acorn.misc

    On 5 Oct 2025 as I do recall,
    Harriet Bazley wrote:

    On 5 Oct 2025 as I do recall,
    Paul Sprangers wrote:

    In article <4a64f8665c.harriet@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    That's quite a common problem in apps that intercept keypresses - so far as I recall you need to explicitly check for F12 and pass it on.

    Okay, KeyMap didn't react on keypresses at all, but now it does, just to pass on F12. Version 1.05 is now ready for download at www.riscos.sprie.nl

    Nice!

    NB if the app doesn't need to react to any keypresses at all, it might
    be simpler just to use the pollmask to mask out the KeyPress Wimp_Poll
    event altogether....
    --
    Harriet Bazley == Loyaulte me lie ==

    Egotist: A person of low taste, more interested in himself than me.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Paul Sprangers@Paul@sprie.nl to comp.sys.acorn.misc on Sun Oct 5 17:54:48 2025
    From Newsgroup: comp.sys.acorn.misc

    In article <fec604675c.admin@bazleyfamily.co.uk>,
    Harriet Bazley <harriet@bazleyfamily.co.uk> wrote:

    (The download seems to have acquired a lot of 'odd' extra keymaps, e.g. 'Test', 'Blank' and 'Woodtype', that you possibly didn't intend to
    include in a distribution version...?)

    NB if the app doesn't need to react to any keypresses at all, it might be simpler just to use the pollmask to mask out the KeyPress Wimp_Poll event altogether....

    All very true. It was all last minute work (concert).

    Paul
    --
    https://riscos.sprie.nl
    --- Synchronet 3.21a-Linux NewsLink 1.2