Re: [OctDev] datestr patch

[Paul K. <pki...@us...>]

On Jun 9, 2005, at 8:58 PM, Keith Goodman wrote:

> Crap. Now I crashed on the format of 'mm'. I would have submitted it 
> all at once had I known.
>
> $ diff -u datestr.m datestr2.m
> --- datestr.m   2005-06-09 17:49:14.677736696 -0700
> +++ datestr2.m  2005-06-09 17:49:31.562169872 -0700
> @@ -25,7 +25,7 @@
>   ## @item  2 @tab mm/dd/yy               @tab 09/07/00
>   ## @item  3 @tab mmm                    @tab Sep
>  ## @item  4 @tab m                      @tab S
> -## @item  5 @tab mm                     @tab 9
> +## @item  5 @tab mm                     @tab 09
>  ## @item  6 @tab mm/dd                  @tab 09/07
>  ## @item  7 @tab dd                     @tab 07
>   ## @item  8 @tab ddd                    @tab Thu
> @@ -124,7 +124,7 @@
>        case { 4, 'm' }
>         str = sprintf("%s",__month_names(M,1));
>        case { 5, 'mm' }
> -       str = sprintf("%d",M);
> +       str = sprintf("%02d",M);
>        case { 6, 'mm/dd' }
>         str = sprintf("%02d/%02d",M,D);
>        case { 7, 'dd' }

Applied.

Note that the 5.2 documentation on the web gives the example of '3' for 
format 'mm'.  Was the documentation wrong or has the definition 
changed?  Is there any way we can support scripts written for older 
versions?  Even if there were, should we bother?

Maintaining compatibility with a language which doesn't even stay 
compatible with itself is frustrating.

- Paul