[Harbour] Why $ can't be "included" anymore?

Szakáts Viktor harbour.01 at syenar.hu
Sun Oct 7 16:57:15 GMT 2007


>> Yes, like the default behaviour of autoadd. Now it is.
>
> And why 1 $ { 1,2 } can't be also?
>
>> Plus, you are talking about _one_ feature (which has
>> a cosmetic significance only IMO).
>
> I don't complain about hb_enumindex() since it can be fixed at link
> time but every $ that we'll miss will be found by the users at RT. But
> the main point is another I don't see the reason to go back to ascan
> for such a simple thing.

That is exactly why I'm strongly against it. I can only repeat
myself: Extending features in such an _invisible_ way _is_ dangerous.
Plus it breaks compatibility, too.

--
local a := {}
? 1 $ a
--

Will give different results in CA-Clipper 5.2 and in Harbour.

Easy to find it in the source? No, it's the most annoying
thing to find, as you've just said.

> Look at TIP for example. Przemyslaw ported it from xHarbour with the
> minimal necessary changes. Then came someone that "changed" it and
> broke the ftp. The problem was tracked and fixed by Patrick Mast in
> xHarbour and I merged ( locally ) his fix but in the meanwhile someone
> who "don't use it" changed it adding "please test".
> Why should I retest my apps every time a developer that don't use some
> code want to rewrite it with a "better syntax".

That's what I was referring to when I told to the list that
I didn't want to "step on anybody's toes". Apparently, it was
your toe.

To me though it looked like the group seemed to favour portable
code over the xhb lib (or HB_COMPAT_XHB) requirement.

If you have a fix for a problem, you could help by making
it public. Just an idea.

> Another example today contrib/gd does not work anymore since it check
> for HB_CONTRIB_XHB.

Someone will need to fix this then, this won't be the first
time such thing happened in Harbour.

Maybe you already have a fix for it? How about publishing it?

> I'm not complaining about mistakes, I do many of them every day, but
> this tells me that contrib is indeed second class code and everyone
> think it can be changed at will without any test.

"Second class" by your definition. Who told you that?

This is an open source project, every contrib and every developer
has the right to create as high quality code as he/she wishes,
_even in contrib_.

If you find that some areas are not developed enough, you
can jump in _anytime_ and fix those problems. In contrib you
even have an almost full freedom to do what you want.

Did you try to do this yet?

Ppl in Harbour have contributed and are contributing huge amounts
of valuable resources from their own, for no direct return.
Understandably, they will first deal with those things which
are really important for them. If this doesn't happen to be
inline with your own priorities, well, you can always contribute.

> So this length discussion is not about "$" but about Harbour future  
> directions.

Do you want to change Harbour's direction towards the xhb direction?

If so, I think you're simply on the wrong mailing list.

> I am developing software since 1980 and I've seen so many things that
> usually I prefer to avoid such discussions but this time is too
> important and I cannot give up or at least I'll fight as long as I
> can.

However you decide Lorenzo.

I really miss that you don't seem to reflect on any of the
positive suggestions others and I have told you, instead you're
doing your fight for the "$" operator, after admitting that you
didn't even follow the latest motives and news in Harbour.

Please reread my suggestions, reread the mails from the past
few weeks (if you haven't done so so far), maybe try those
suggestions and also understand that this is not xhb.

Brgds,
Viktor



More information about the Harbour mailing list