Re: [Bug-readline] Forwarding input from ncurses to readline

[Ulf Magnusson]

On Tue, Feb 17, 2015 at 10:56 PM, Ulf Magnusson <address@hidden> wrote:
> On Tue, Feb 17, 2015 at 10:48 PM, Chet Ramey <address@hidden> wrote:
>> On 2/17/15 4:25 PM, Ulf Magnusson wrote:
>>
>>> Is the following comment in terminal.c outdated by the way?
>>>
>>> /* If we're being compiled as part of bash, set the environment
>>>    variables $LINES and $COLUMNS to new values.  Otherwise, just
>>>    do a pair of putenv () or setenv () calls. */
>>> if (rl_change_environment)
>>>   sh_set_lines_and_columns (_rl_screenheight, _rl_screenwidth);
>>>
>>> It doesn't seem to matter whether readline is compiled as part of bash.
>>
>> It does, actually.  The trick is the traditional Unix linker behavior.
>> There is a source file in the readline library that provides a set of
>> functions whose names begin with sh_ (shell.o):
>>
>> sh_single_quote (string)
>> sh_set_lines_and_columns (lines, cols)
>> sh_get_env_value (varname)
>> sh_get_home_dir ()
>> sh_unset_nodelay_mode (fd)
>>
>> When readline is linked as part of bash, those are resolved from the
>> bash binary, since those names are already known to the linker when it
>> processes libreadline.a.  All of the functions in that file are resolved
>> from bash, so the file is not linked in.  When readline is not being
>> compiled as part of bash, those functions are not known when processing
>> libreadline.a, and are resolved using the built-in readline versions, so
>> the symbols in shell.o are used.
>
> Ah, that clarifies it. :)
>
>> If you don't want readline to modify $LINES and $COLUMNS, set
>> rl_change_environment to 0.
>>
>> Chet
>
> Yeah, I'll do that.
>
> Thanks,
> Ulf

Re. rl_change_environment, there's this in the source:

/* If this is non-zero, readline will set LINES and COLUMNS in the
   environment when it handles SIGWINCH. */
int rl_change_environment = 1;

However, afaics readline does not update LINES and COLUMNS in the
SIGWINCH handler, at least not in the non-bash case. Is there some
motivation for this (besides being a bit tricky due to putenv/setenv not being
async-signal-safe)? The comment seems a little misleading either way. :)

(I've set rl_catch_signals and rl_catch_sigwinch to 0 already to decouple
things further from ncurses, but got curious.)

/Ulf