Re: [Nano-devel] .nano_history loading

[David Lawrence Ramsey]

--- David Benbennick <address@hidden> wrote:
>Second try.  There was a double-free() (because of the static 
>variable).
>  
>New version, I made a global homedir variable, to make sure it doesn't 
>change while nano is running.  Also, an error reading .nano_history 
>used to be invisibly reported.
>
>Finally, while translation is an issue: why not leave the rcfile_error 
>messages without final \n's, and just write a new line in itself?  The 
>patch does that.  That way translators don't have to make all those 
>changes.

Sorry for the lack of responses lately.  I've been on a slower than 
usual connection lately (i.e, dialup), and I've been focusing more on 
keeping up with the patch queue under it.

In any case, I've been merging the bulk of your previous patches, as you 
could probably tell from your having to resync them with CVS; they've 
definitely been useful.

As for your latest patch, I've merged the rcfile_error() message changes 
and the prompting for when there's an error reading .nano_history, plus 
a few other miscellaneous bits, all of which seem good to me, but the 
actual overhaul of the history-reading functions seems a bit too 
intrusive at the moment.  Specifically, the reliance on getline() is a 
problem.  As you said before, there probably should be a version of it 
for non-GNU machines, and I'd agree with that, mainly because not having 
one would give e.g. the BSD users problems.  At the moment, though, 
there isn't one, and I don't want to delay 1.3.4 anymore just to get one 
written, especially as 1.3.4 is supposed to be the relatively stable 
release that 1.3.3 should have been.

I'll definitely look into this again once 1.3.4-cvs is out, although I'm 
mainly planning to focus on UTF-8 support for 1.3.5.