apparently unicode support was lost with sbcl 2.0.3

[cage2]

Hi,

I do not know if just me but after upgraded to sbcl 2.0.3 the support for widechar seems gone and, fir unicode characters, nothing is printed instead; to reproduce just run the test ut01.

Bye!
C.

[McParen]

Thanks, I have not yet tried 2.0.3, I will do that next. At the moment, I am still trying to find out what caused the unicode background routine in ncurses to stop working (issue #28).

[cage2]

On Mon, Apr 06, 2020 at 11:55:53AM -0700, McParen wrote: Hi!

Thanks, I have not yet tried 2.0.3, I will do that next. At the moment, I am still trying to find out what caused the unicode background routine in ncurses to stop working (issue #28).

Sure, no problem! Let's consider this issue report just as a memo. Bye! C.

[cage2]

On Mon, Apr 06, 2020 at 11:55:53AM -0700, McParen wrote: Hi!

Thanks, I have not yet tried 2.0.3, I will do that next. At the moment, I am still trying to find out what caused the unicode background routine in ncurses to stop working (issue #28).

So, i have investigated a bit and seems that croatoan fell victim of the recent decision to use utf8 as default external format. :) sbcl/sbcl@2454a32 If you take a look at the diff you'll notice that the function 'setup_locale()' was removed aftre the changes mentioned above. Fortunately you already provided the workaround, just define and call, befoe `initscr', the setlocale as defined in example `ut05' of ncurses test and wide char supports came back! :) Of course this is nothing more than a starting point, probably a better solution needs to be planned. Hope this helps! Bye!

[McParen]

just define and call, befoe `initscr', the setlocale

Thanks for finding that out, this essentially solves the issue.

probably a better solution needs to be planned.

Why? The call to setlocale is the recommended way to properly start ncurses. From the man page:

The  library uses the locale which the calling program has initialized.
That is normally done with setlocale:

    setlocale(LC_ALL, "");

If the locale is not initialized, the library assumes  that  characters are
printable  as in ISO-8859-1, to work with certain legacy programs.
You should initialize the locale and not rely on  specific  details  of the
library when the locale has not been setup.

Until now, we have simply relied on SBCL doing this for us by default, now we have to do it manually, as every other ncurses app.

[cage2]

On Tue, Apr 07, 2020 at 10:27:03AM -0700, McParen wrote: Hi McParen!

> just define and call, befoe `initscr', the setlocale Thanks for finding that out, this essentially solves the issue.

Good! Happy to be useful! :)

> probably a better solution needs to be planned. Why? The call to setlocale is the recommended way to properly start ncurses. From the man page:

[...]

Until now, we have simply *relied* on SBCL doing this for us by default, now we have to do it manually, as every other ncurses app.

Actually i thought it was just to simple, the only thing i did was to check runtime.c of SBCL (as suggested by your comment) and paste your function into the example. Seemed just a quick hack to me, turned out was a good solution, instead. I find this funny someway! :D Bye! C.

[McParen]

I've pushed a commit adding a call to setlocale to the initialization of screen objects. For direct ncurses calls, you will have to setlocale manually every time, but that's intended.

For me this solves the issue, can it be closed?

[cage2]

On Wed, Apr 08, 2020 at 12:31:38AM -0700, McParen wrote: Hi!

I've pushed a commit adding a call to setlocale to the initialization of screen objects. For direct ncurses calls, you will have to setlocale manually every time, but that's intended. For me this solves the issue, can it be closed?

Wonderful, works like a charm! Thank you! C.