Re: [PUSHED] vcl/source/window/status.cxx: Translate German comments

On 01/17/2014 12:50 PM, Sven Wehner wrote:
> sorry, I guess I misspoke:
> Those examples, which I listed, are rather obvious problems that aren't translation related.
> So let me rephrase:
>         How to call a developer's attention to small pieces of code?
>         And especially: may I combine that with translation patches?

I did get that (and think Philipp gave a good response in the 
meantime---I'd only add that if you do find something that looks 
genuinely broken, best raise it on the dev ML or IRC channel).

> I thought "OSL_ENSURE(false, …)" would be "problematic", because firstly there is an "OSL_FAIL" definition in 
> "include/osl/diagnose.h", which seems to do exactly that.
> Secondly, the place I saw this (cui/source/tabpages/numpages.cxx line 2791 ff.) encapsulates that 
> statement in an if clause.
> Together with the error message ("cannot happen"), it implies that could be reduced from…
> --- snip ---
>      if (SVX_MAX_NUM <= nLvl)
>      {
>          OSL_ENSURE(false, "cannot happen.");
>          return;
>      }
> --- snap ---
> … to something like this:
> --- snip ---
>          OSL_ENSURE_RETURN(!(SVX_MAX_NUM <= nLvl), "cannot happen.");
> --- snap ---
> (with OSL_ENSURE_RETURN being a OSL_ENSURE including a return statement.)
>
> If OSL_ENSURE is deprecated anyway, I guess this is irrelevant anyway.

That specific case should probably be rewritten as either

  if (nLvl >= SVX_MAX_NUM) {
     SAL_WARN("cui.tabpages", "should never happen");
     return;
  }

or

  assert(nLvl < SVX_MAX_NUM)

depending on context (see 
<https://wiki.documentfoundation.org/Development/GeneralProgrammingGuidelines#Assertions_and_Logging> 
for details), but there is already EasyHack 
<https://bugs.freedesktop.org/show_bug.cgi?id=43157> "Clean up 
OSL_ASSERT, DBG_ASSERT, etc.," so no need to call developers' attention 
on OSL_ENSURE(false,...) in general.

Stephan