|
| From: | Jeff Abrahamson |
| Subject: | Re: [RP] select patch |
| Date: | Wed, 25 Jun 2014 22:52:34 +0100 |
It looks very good, thanks!Want to update the docs? ;-)Disclaimer: I haven't tested this myself.+Jérémie, since I think he may be the only one with commit privileges.
On 24 June 2014 14:22, Johannes Altmanninger <address@hidden> wrote:
Thanks for your help! There is a whole lot to learn for me..
I corrected most issues you pointed out, this keeps getting better ;)
I changed the default behavior to prefer unselected windows, the option to change it is still there. Most users would probably not care too much about this I guess.
Now there is a function called get_compare_string_function() which is used in select_name(). It takes the match_type bitmask and the address of compare_length.
I am not sure how I would implement a function that does not take the match type as argument.
I noticed that there are 8 occurrences of str_comp(), is there a reason why this is not replaced by `!strncasecmp()`?
On 06/23/2014 04:30 PM, Jeff Abrahamson wrote:
> Funny how documentation is always the trailing bit. ;-)
> This is looking very nice. Thanks for all your work on this!
> A couple (easy) comments:
> Just delete the stuff you're deleting, don't comment it out. Git remembers what was there before, and it just makes your diffs longer. I know it's useful while you're developing.
> I'm acutely aware that I'm the guy who called for the optional behavior -- but the more I think about it, the more I think the default should be the behavior you propose. It is a very good idea.
> On that subject, the name "SELECT_SIMPLE" seems prejudiced by the old behavior and otherwise devoid of meaning. What do you think of SELECT_OK_SELECTED ?
> When you define MAX_WINDOW_NAME_LENGTH, I think the value is arbitrary, just long enough to distinguish between valid and invalid strings in reasonable time. Perhaps include a comment to that effect so that future readers don't pause to wonder on the significance of that number.
> Personal tick: Functions are most readable below ten or so lines. By the time I can't see the whole function in my editor, I have more trouble understanding as I read. In that light, and without remarking on other examples of such, perhaps find_window_name() could easily be broken into a helper function or two (with local linkage), such as a
>
> compare_window_name_function window_name_matcher(size_t& compare_length).
>
> Jeff Abrahamson
> +33 6 24 40 01 57 <-- brièvement indisponible le 4 juillet
> +44 7920 594 255 <-- will change 18 July
>
> http://jeff.purple.com/
> http://blog.purple.com/jeff/
| [Prev in Thread] | Current Thread | [Next in Thread] |