insstr (from ncursesw) UTF-8 issue

[Ciprian Dorin, Craciun]

    Hello all!

    I've discovered a problem (I think it's a bug, but maybe I'm doing
something wrong) about the behavior of the insstr procedure in the
context of UTF-8 encodings. (I'm using the ncursesw library.)

    To keep the email short, I've attached a tar file that contains
all the necessary code to reproduce the bug. (Both a Python version
(from which I've started) and a C version.)
    * context.txt contains the current versions of ncursesw library,
and the result of the linking of my application;
    * bug.c and bug.py contain the sourcecode;
    * bug.c.bash and bug.py.bash the commands used to run the code;

    In summary the following code does not work correctly:
...
        mvinsstr (4, 0, "|aaist|"); // the correct expected order
(without special characters)
        mvinsstr (5, 0, "|\xc4\x83\xc3\xa2\xc3\xae\xc5\x9f\xc5\xa3|");
// the wrong order (with special characters)
...

    It should print the characters |ăâîşţ| (some Romanian diacritics)
(as in |aaist|), but instead it prints ||ţşîâă. It seems it sorts the
characters... Now the addstr works as expected.

    Any problems on my side? Any solutions?

    Thanks,
    Ciprian Craciun.


_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Ciprian Dorin, Craciun]

    Any updates on the instr issue?


On Tue, Sep 8, 2009 at 3:49 AM, Thomas Dickey <dickey@...> wrote:

> On Mon, 7 Sep 2009, Ciprian Dorin, Craciun wrote:
>
>>   Hello all!
>>
>>   I've discovered a problem (I think it's a bug, but maybe I'm doing
>> something wrong) about the behavior of the insstr procedure in the
>> context of UTF-8 encodings. (I'm using the ncursesw library.)
>>
>>   To keep the email short, I've attached a tar file that contains
>> all the necessary code to reproduce the bug. (Both a Python version
>> (from which I've started) and a C version.)
>>   * context.txt contains the current versions of ncursesw library,
>> and the result of the linking of my application;
>>   * bug.c and bug.py contain the sourcecode;
>>   * bug.c.bash and bug.py.bash the commands used to run the code;
>>
>>   In summary the following code does not work correctly:
>> ....
>>       mvinsstr (4, 0, "|aaist|"); // the correct expected order
>> (without special characters)
>>       mvinsstr (5, 0, "|\xc4\x83\xc3\xa2\xc3\xae\xc5\x9f\xc5\xa3|");
>> // the wrong order (with special characters)
>> ....
>>
>>   It should print the characters |ăâîşţ| (some Romanian diacritics)
>> (as in |aaist|), but instead it prints ||ţşîâă. It seems it sorts the
>> characters... Now the addstr works as expected.
>
> Offhand, it looks like a bug in ncurses - I'll check/see what the fix would
> be...
>
> (ins_wstr and related functions should work, though they're more cumbersome
> to use).
>
> --
> Thomas E. Dickey
> http://invisible-island.net
> ftp://invisible-island.net

_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Thomas Dickey-2]

On Thu, 15 Oct 2009, Ciprian Dorin, Craciun wrote:

>    Any updates on the instr issue?

I made a fix here:

20090912
         + add some test programs (and make these use the same special keys
           by sharing linedata.h functions):
           test/test_addstr.c
           test/test_addwstr.c
           test/test_addchstr.c
           test/test_add_wchstr.c
         + correct internal _nc_insert_ch() to use _nc_insert_wch() when
           inserting wide characters, since the wins_wch() function that it used
           did not update the cursor position (report by Ciprian Craciun).

(it seems to work; I'll be out of town this Saturday and won't have time
for more than a small fix or two before then).

>
>
> On Tue, Sep 8, 2009 at 3:49 AM, Thomas Dickey <dickey@...> wrote:
>> On Mon, 7 Sep 2009, Ciprian Dorin, Craciun wrote:
>>
>>>   Hello all!
>>>
>>>   I've discovered a problem (I think it's a bug, but maybe I'm doing
>>> something wrong) about the behavior of the insstr procedure in the
>>> context of UTF-8 encodings. (I'm using the ncursesw library.)
>>>
>>>   To keep the email short, I've attached a tar file that contains
>>> all the necessary code to reproduce the bug. (Both a Python version
>>> (from which I've started) and a C version.)
>>>   * context.txt contains the current versions of ncursesw library,
>>> and the result of the linking of my application;
>>>   * bug.c and bug.py contain the sourcecode;
>>>   * bug.c.bash and bug.py.bash the commands used to run the code;
>>>
>>>   In summary the following code does not work correctly:
>>> ....
>>>       mvinsstr (4, 0, "|aaist|"); // the correct expected order
>>> (without special characters)
>>>       mvinsstr (5, 0, "|\xc4\x83\xc3\xa2\xc3\xae\xc5\x9f\xc5\xa3|");
>>> // the wrong order (with special characters)
>>> ....
>>>
>>>   It should print the characters |ăâîşţ| (some Romanian diacritics)
>>> (as in |aaist|), but instead it prints ||ţşîâă. It seems it sorts the
>>> characters... Now the addstr works as expected.
>>
>> Offhand, it looks like a bug in ncurses - I'll check/see what the fix would
>> be...
>>
>> (ins_wstr and related functions should work, though they're more cumbersome
>> to use).
>>
>> --
>> Thomas E. Dickey
>> http://invisible-island.net
>> ftp://invisible-island.net
>

--
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net
_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses

[Ciprian Dorin, Craciun]

On Fri, Oct 16, 2009 at 2:34 AM, Thomas Dickey <dickey@...> wrote:

> On Thu, 15 Oct 2009, Ciprian Dorin, Craciun wrote:
>
>>   Any updates on the instr issue?
>
> I made a fix here:
>
> 20090912
>        + add some test programs (and make these use the same special keys
>          by sharing linedata.h functions):
>          test/test_addstr.c
>          test/test_addwstr.c
>          test/test_addchstr.c
>          test/test_add_wchstr.c
>        + correct internal _nc_insert_ch() to use _nc_insert_wch() when
>          inserting wide characters, since the wins_wch() function that it
> used
>          did not update the cursor position (report by Ciprian Craciun).
>
> (it seems to work; I'll be out of town this Saturday and won't have time
> for more than a small fix or two before then).
>
>>
>>
>> On Tue, Sep 8, 2009 at 3:49 AM, Thomas Dickey <dickey@...> wrote:
>>>
>>> On Mon, 7 Sep 2009, Ciprian Dorin, Craciun wrote:
>>>
>>>>   Hello all!
>>>>
>>>>   I've discovered a problem (I think it's a bug, but maybe I'm doing
>>>> something wrong) about the behavior of the insstr procedure in the
>>>> context of UTF-8 encodings. (I'm using the ncursesw library.)
>>>>
>>>>   To keep the email short, I've attached a tar file that contains
>>>> all the necessary code to reproduce the bug. (Both a Python version
>>>> (from which I've started) and a C version.)
>>>>   * context.txt contains the current versions of ncursesw library,
>>>> and the result of the linking of my application;
>>>>   * bug.c and bug.py contain the sourcecode;
>>>>   * bug.c.bash and bug.py.bash the commands used to run the code;
>>>>
>>>>   In summary the following code does not work correctly:
>>>> ....
>>>>       mvinsstr (4, 0, "|aaist|"); // the correct expected order
>>>> (without special characters)
>>>>       mvinsstr (5, 0, "|\xc4\x83\xc3\xa2\xc3\xae\xc5\x9f\xc5\xa3|");
>>>> // the wrong order (with special characters)
>>>> ....
>>>>
>>>>   It should print the characters |ăâîşţ| (some Romanian diacritics)
>>>> (as in |aaist|), but instead it prints ||ţşîâă. It seems it sorts the
>>>> characters... Now the addstr works as expected.
>>>
>>> Offhand, it looks like a bug in ncurses - I'll check/see what the fix
>>> would
>>> be...
>>>
>>> (ins_wstr and related functions should work, though they're more
>>> cumbersome
>>> to use).
>>>
>>> --
>>> Thomas E. Dickey
>>> http://invisible-island.net
>>> ftp://invisible-island.net
>>
>
> --
> Thomas E. Dickey
> http://invisible-island.net
> ftp://invisible-island.net

    Yes it works. Thanks!


_______________________________________________
Bug-ncurses mailing list
Bug-ncurses@...
http://lists.gnu.org/mailman/listinfo/bug-ncurses