I doubt "while-else" is a variant, due to dangling "else".

How will be compiled, e.g.
if something then
  while not_done_yet do
    the_work
else // obviously valid, and expected to be a part of "if-then-else"
  something_else_happened;

---
WBR, Max.

That is the meaning of "However, both syntax would cause code incompatibility." and is already in my very first reply to OP.

So While... else or While .. Otherwise will never make it in the compiler.

Michael.

Thanks for that.

This is s typical example of one of those things that just don't make any sense in the first place.

Either the while loop is executed or it isn't, depending in the expression. I don't see an actual use case for any else/otherwise extension to it...

Ralf

---
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

You probably want to reread python while-else implementation. ( https://docs.python.org/2/reference/compound_stmts.html) "Else" becomes sort of "part of the loop". Thus if you break out of the loop, "else" would not be executed.
However, if no break occurs "else" part would be executed.

It might be a rare case where it's needed. (I cannot think of any), but to achieve exactly the same functionality in Pascal either of two options should be used.

1) make an extra check if break occurred.

breakflag:=false;
while cond do
  if someothercond then begin
    breakflag : =true;
    break;
  end;
if not breakflag then
  while_else_code

2) use goto! :)

label postElseLabel

while cond do
  if someothercond then begin
     goto  postElseLabel
  end;

while_else_code
:postElseLabel

thanks,
Dmitry

Sorry, we are programming here in Pascal, not Python. I could give a rodent's posterior about what Python implements.

People need to properly design their software, not just hacking away. Then they do not need to come up with whacko stop-gaps to fix what could have been done properly in the first place...

Ralf

---
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

Well, from fpc-devel list perspective your only hope is that any new feature is added via MODESWITCH. Thus any existing unit would be kept clean. You would also be able to keep your libraries clean, by not allowing using of a certain modeswitches.

However, if you look at the recent (for the past few months) discussions, you might spot that the number of requests for new features (from recent Delphi and other languages) is growing. Constantly.

In this particular case it's even worse (actually it's better, from fpc-devel POV), since the patch has been provided.

The next step would probably be controlled "break", where a user would be able to specify how many nested loops needed to broken from.

thanks,
Dmitry

The only possible use would be to specify a finalisation statement which was only executed if the while loop cycled at least once, or possibly one which was executed if the loop never cycled.

I can't think of a use case either. And just because another language provides such a solution, does not imply that FPC should. As to achieving the functionality in FPC:

if cond then
   while cond do
     foo
else
   bar;

should suffice; Breakflags and gotos not required!

Am 12.10.2015 20:02 schrieb "Dmitry Boyarintsev" <

rodent's posterior about what Python implements.

> > People need to properly design their software, not just hacking away.

Then they do not need to come up with whacko stop-gaps to fix what could have been done properly in the first place...

Well, from fpc-devel list perspective your only hope is that any new

feature is added via MODESWITCH. Thus any existing unit would be kept clean. You would also be able to keep your libraries clean, by not allowing using of a certain modeswitches.

Since one point of this discussion is to implement in a way that it doesn't break anything, it wouldn't be implemented.behind a modeswitch probably.

However, if you look at the recent (for the past few months) discussions,

you might spot that the number of requests for new features (from recent Delphi and other languages) is growing. Constantly.

True...

In this particular case it's even worse (actually it's better, from

fpc-devel POV), since the patch has been provided.

However we did nowhere state that we'd incorporate the feature. I only said that I'd definitely look at it to check whether the work in principle is good.

The next step would probably be controlled "break", where a user would be

able to specify how many nested loops needed to broken from.

That would be more useful than the while-else though :P

Regards,
Sven

Yes, I have seen a lot of those requests and pretty much none of them made any sense to me. As I already stated, it always seems to me that someone wants to have something implemented that they have seen in a different language, which might have a different approach to programming to begin with or they "think" they need a specific feature to get them out of a pickle they got themselves into because they are just hacking away instead of properly designing their programs. Neither of those case are IMHO valid to mess around with a working compiler. If someone is so fond of the features that Python/Haskell/Erland/<your favorite 'Cool-boy' programming language de jour> provides, then they should use that to solve their programming issues. They apparently don't understand how to use Pascal...

Ralf

---
This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus

I guess I'm confused. What is the difference between:

while true do this then that

and:

while true do this
that

And what is better about:

while true do this otherwise/else that

and:

if true
while true do this
else
that

The latter may have a difference if the thing that returns true for the "if" could be false by the time the "while" is executed, I suppose.

Jeff.

that just doesn't make grammatical sense...

foo := 0;
while foo < 100 do
   inc(foo);
else
   dec(foo);
end

so, if foo is less than 100 we inc(foo)... if foo is not less than 100 we dec(foo)... that's the only way that an "else" can be read... either the boolean is true or /else/ it is false... only if it is false can the else portion be executed...

(#2) but then the question is does dec(foo) execute only once or does it execute as long as foo is greater than 100??

what i seem to see is pure laziness of a sort... folks too lazy to write an "if" or "if/else" statement and put the while inside...

foo := 0;
if foo < 100 then
   while foo < 100 do
     inc(foo);
else
   dec(foo);

OR depending on the answer of (#2) above...

foo := 0;
if foo < 100 then
   while foo < 100 do
     inc(foo);
   while foo > 100 do
     dec(foo);

or maybe this is the actual intent?

foo := 0;
if foo < 100 then
   while foo < 100 do
     inc(foo);
if foo > 100 then
   dec(foo);

seems like a pretty tough way to have foo equal to 99, eh? ;)

[...]

2) use goto! :)

neither is needed at all in most cases and then not if they have proper logic in place ;)

Dmitry Boyarintsev wrote

The next step would probably be controlled "break", where a user would be able to specify how many nested loops needed to broken from.

ROTFLMAO! if you need or desire something like that then set a breakcounter and break... in the next outer block, check breakcounter to see if you need to break again... if so, dec(breakcounter) and break... then check it again in the next outer block... and so on and so on until breakcounter=0... why should the compiler have to do your logic work for you? ;)

How would you know if a nested loop finished properly or broke out? ...Just imagine yourself a 4 nested for loops and you need to break out from the 4th to the 1st?

thanks,
Dmitry

by checking the value that caused the break ;) deity knows i've done it many times before back in the TP/BP 6&7 days... i did it exactly as described, too... we had to do it that way as there is/was no other way to do it ;)

by checking the value that caused the break ;) deity knows i've done it many times before back in the TP/BP 6&7 days... i did it exactly as described, too... we had to do it that way as there is/was no other way to do it ;)

That's exactly the thing. You didn't have a cross-loop break, that's why you've to have some guaranteed values that would allow you to verify if a loop was broken or not.

If it was broken then you'd set any additional conditions (if needed) to break out another loop and pass it above.

In the end you end-up having additional break condition checks in each loop. Just by the fact, that a language doesn't provide you with any other means to do it less complex.

I'd also need to note, that no other language, to my knowledge, have cross-loop breaks anyway :)
Good-old language (Pascal / C /C++) still have goto though... but we won't use it, right?

thanks,
Dmitry

yeah... i implemented my needed logic for that type of scenario... just like any good programmer can/will/should do...

ok... so instead we'll have the compiler perform my business logic for me? no thanks... the compiler has enough to do already ;)

I'd also need to note, that no other language, to my knowledge, have cross-loop breaks anyway :) Good-old language (Pascal / C /C++) still have goto though... but we won't use it, right?

that's right... i think i may have used goto once in 30 years or however long it has been available in pascal... i started self-taught with basic... goto and gosub all over the place... tough as hades to follow the code and no apparent logical layout... i was ecstatic when a friend turned me on to pascal and asm... i dropped basic like a hot potato[e?] and never looked back... especially never looking back at goto ;) ;) ;)

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, 12 Oct 2015 15:11:18 -0400, Wkitty42 ()
wrote about "Re: [fpc-devel] Fwd: While - Otherwise Statement" (in ):

Hi Mark,

You might have seen me write things like this in the old OS2PROG echo of Fidonet some 20+ years ago. ... :-)

The only language I know that offers that level of control is PL/I, where the break statement is coded as LEAVE. It is handled by labelling the loop control statement and coding the required label in the LEAVE statement. For example:

   loop_1: DO i = 1 TO m;
      loop_2 : DO j = 1 TO n;
         loop_3: DO k = 1 TO p;
            ...
            IF some_condition THEH
               LEAVE loop_3;
            ...
            IF another_condition THEH
               LEAVE loop_2;
            ...
            IF some_other_condition THEH
               LEAVE loop_1;
            ...
         END loop_3;
      END loop_2;
   END loop_1;

If the label is omitted then the immediately containing loop is left.

This allows any of the nested loops to be escaped, potentially all the way out. I guess the corresponding syntax in a Pascal-esque style would be:

   loop_1: FOR i := 1 TO m DO
   BEGIN
      loop_2: FOR j := 1 TO n DO
      BEGIN
         loop_3: FOR k := 1 TO p DO
         BEGIN
            ...
            IF (some_condition) THEN
               BREAK loop_1;
            ...
            IF (another_condition) THEN
               BREAK loop_2;
            ...
            IF (some_other_condition) THEN
               BREAK loop_3;
         END
      END
   END;

If we're being strictly Pascal, the labels would need to be declared at the head of the procedure -- and if we're really strict they should be numeric.
- --
Regards,

Dave [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iEYEARECAAYFAlYcMQMACgkQogYgcI4W/5Sl3wCgtG87fG2aCEuxkVCUrBCmxwCm fSkAoNgXwMSNA/R4AKjDEZsNktglOqvF
=m3vh
-----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, 12 Oct 2015 23:15:31 +0100, David W Noon
() wrote about "Re: [fpc-devel] Fwd: While
- - Otherwise Statement" (in ):

[snip]

IF some_condition THEH IF another_condition THEH IF some_other_condition THEH

These should all end in "THEN", not "THEH".
- --
Regards,

Dave [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iEYEARECAAYFAlYcM7sACgkQogYgcI4W/5SQagCfV6i2NrBee3h76kegPVARLq5V j7sAoNNPDGBtwsEWqdrrRiLylwoJkH7k
=ZVGL
-----END PGP SIGNATURE-----

hiya dave!! it has been a long time, my friend!!

You might have seen me write things like this in the old OS2PROG echo of Fidonet some 20+ years ago. ... :-)

oh yeah! that echo still exists, too... but it only contains monthly moderator postings these days... it is ready for you any time you want to join back in... nntp access is available via some systems but the dawg grows older with each passing year :?

The only language I know that offers that level of control is PL/I, where the break statement is coded as LEAVE. It is handled by labelling the loop control statement and coding the required label in the LEAVE statement. For example:

that looks very much like what some would consider goto statements... does the leave return to the top of the previous loop or does it drop to the next statement in the previous loop?

If the label is omitted then the immediately containing loop is left.

"immediately containing loop" meaning loop_3 if we're in loop_3?

This allows any of the nested loops to be escaped, potentially all the way out. I guess the corresponding syntax in a Pascal-esque style would be:

that does look interesting for breaks and i can see how it may be beneficial and save some bit of coding but then those of us who have been around a while know how to work our way back out when necessary, right? ;)

If we're being strictly Pascal, the labels would need to be declared at the head of the procedure -- and if we're really strict they should be numeric.

yup! and i did catch the THEH->THEN oops... no problems there O:)

If I'm not missing something, Java implements 'break' just like this. I'd like such enhancement, but right now is there any reason to avoid 'goto' besides the bad karma?

Regards,
Flávio

The intrafunctional gotos not really. But the interfunctional ones (modeswitch nonlocalgoto in 2.7.1+) might mess with managed types ;)

Regards,
Sven

Sven Barth wrote

The intrafunctional gotos not really. But the interfunctional ones (modeswitch nonlocalgoto in 2.7.1+) might mess with managed types ;)

The compiler should give an error if you try to "goto" out of or into a frame than needs initialisation/finalisation.

Jonas

frame than needs initialisation/finalisation.

Ah, ok, so we're already checking that. Good ^^
Then it's mostly the "bad karma" of goto I'd say ;)

Regards,
Sven

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, 12 Oct 2015 22:34:03 -0300, Flávio Etrusco () wrote about "Re: [fpc-devel] Fwd: While -
Otherwise Statement" (in
):

On Mon, Oct 12, 2015 at 7:15 PM, David W Noon wrote:

[snip]

Last time I coded any significant amount of Java it still used the C/C++ syntax -- but that was 10 or 11 years ago.

I'd like such enhancement, but right now is there any reason to avoid 'goto' besides the bad karma?

I view it largely as "the curse of Edsgar Dijkstra". A very short range goto should not be a real problem.
- --
Regards,

Dave [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iEYEARECAAYFAlYdI7gACgkQogYgcI4W/5QdGACgt7fbezm0wbDDASxjINbUI/tP rN0AnRSl7Ja7Sk3nIRGPqwQQUvAS0AqJ
=AQvj
-----END PGP SIGNATURE-----

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mon, 12 Oct 2015 21:32:07 -0400, Wkitty42 ()
wrote about "Re: [fpc-devel] Fwd: While - Otherwise Statement" (in ):

[snip]

that looks very much like what some would consider goto statements... does the leave return to the top of the previous loop or does it drop to the next statement in the previous loop?

It leaves the loop at the END statement that matches the DO statement as labelled. So it goes to the next statement in the surrounding loop.

> If the label is omitted then the immediately containing loop is > left. "immediately containing loop" meaning loop_3 if we're in loop_3?

Correct. It's the same as break.

yup! and i did catch the THEH->THEN oops... no problems there O:)

Ageing eyes, I'm afraid. When I remove my contact lenses at night I have very distorted corneas (keratoconus) and cannot easily distinguish an H from an N. ... :-(
- --
Regards,

Dave [RLU #314465]
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* (David W Noon)
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iEYEARECAAYFAlYdJV4ACgkQogYgcI4W/5QZKgCaA7lhDyhW/D0E+ATdBhrkzyIM YrsAmwYelE+svkFxHDZvKLP3AJ/HZ8DN
=WNCc
-----END PGP SIGNATURE-----