OpenEuphoria

2. Re: Euphoria 4.0.1

• Posted by jaygade Mar 12, 2011
• 2765 views

I don't see any major issues unless any particular tickets need to be worked on. What date do you see for release?

Right now (12 March 2011) my 4.0 branch has one test fail (t_eutest.e) and my main branch has 3 tests fail (translated t_bitwise.e, t_eutest.e, t_pipeio.e).

4. Re: Euphoria 4.0.1

• Posted by jaygade Mar 12, 2011
• 2743 views

Example:

include std/sequence.e 
sequence A 
 
A = series(1, 1, 10) 
 
? A 
 
{1, 2, 3, 4, 5, 6, 7, 8, 9, 10} 

Reference: http://openeuphoria.org/docs/std_sequence.html#_2498_series

Example:

? A / 2 
 
{0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5} 

This has ALWAYS been a part of Euphoria. It's a long standing-feature and a strong selling point of the language, otherwise Euphoria would just be "Pascal Lite". Reference: http://openeuphoria.org/docs/lang_def.html#_8_operationsonsequences

I'm not sure I understand this last bit, but isn't it just:

object B = 43.89, C 
 
C = {B} 
 
? B 
? C 
 
43.89 
{43.89} 

You can always use atom() or sequence() to determine if something is an atom or a sequence. Reference: http://openeuphoria.org/docs/lang_decl.html#_118_atom
http://openeuphoria.org/docs/lang_decl.html#_119_sequence (Hmm. That could stand to be re-written to describe better what the sequence() function does.)

6. Re: Euphoria 4.0.1

• Posted by Vinoba Mar 13, 2011
• 2660 views

 
-- Start of test code stest.ex 
 
without type_check 
 
without warning 
 
include std/console.e 
 
include machine.e 
 
include sort.e 
 
include misc.e 
 
sequence spurchages = { "Screws" , "Nails" , "Rivets" } 
 
sequence spr = { 7.34,  2.54,  6.72 } 
 
 sequence squan = { 12, 20, 24 } 
 
-- atom squan = 12 
 
-- sequence squan = 12 
 
-- integer squan = 12 
 
-- sequence squan = { 12, 12, 12 } 
 
atom ttotal = 0 
 
sequence sdetails 
sdetails = invoice ( spr , squan ) 
 
-- Now go and print invoice using spr, squan, sdetails and ttotal  
 
function invoice (sequence sprice , sequence squantity ) 
 
	-- Multiply each element of sequence  sprice with each element of squantity and return in ttotal 
 
      integer isize = length ( sprice ) 
 
      sequence sitemcost = repeat (0 , isize) 
 
      for i = 1 to isize do 
 
        sitemcost [ i ] = sprice [ i ] * squantity [ i ] 
 
        ttotal = ttotal + sitemcost [ i ]  
 
	end for 
 
      return sitemcost 
 
end function 
 
-- End of test code stest.ex 
 

In this example, if I use the quantities in squan as a full matching length sequence, I am OK. But look at the alternatives I have commented out.

But supposing I want to use a derived amount as a result of some other calculations or simply if the quantities for all items purchased is 12, I have to use 12, 12, 12.

Thus the facility of one value being used against many in Math operations is not properly usable. I can use atoms but even then within the function, I have to do a direct

sitemcost = sprice * squantity

There is no comfortable way of doing the invoice presuming that a single value for quantity or price would be acceptable within the function off and on as the situation arises. Also if the invoicing function has to be used for a single item purchased the function will not work comfortably.

8. Re: Euphoria 4.0.1

• Posted by mattlewis (admin) Mar 13, 2011
• 2633 views

I'm not sure what the issue is here. You gave the solution that accomplishes what you appear to want. You could also use math:sum:

include std/math.e 
function invoice (sequence sprice , object squantity )  
	-- Multiply each element of sequence  sprice with each element of squantity and return in ttotal 
	sequence sitemcost = sprice * squantity 
	ttotal = sum( sitemcost & ttotal ) 
	return sitemcost  
end function  

Now, you end up with a sequence of totals, plus the total. For a single item, you simply have a sequence with a single price, or you do something like this (you'll find this idiom in a lot of euphoria code:

include std/math.e 
function invoice (object sprice , object squantity )  
	if atom( sprice ) then 
		sprice = {sprice) 
	end if 
	-- Multiply each element of sequence  sprice with each element of squantity and return in ttotal  
	sequence sitemcost = sprice * squantity 
	ttotal = sum( sitemcost & ttotal ) 
	return sitemcost  
end function  

Did I miss something else that you wanted to be able to do? Or why this isn't an acceptable solution?

Matt

10. Re: Euphoria 4.0.1

• Posted by mattlewis (admin) Mar 13, 2011
• 2605 views

You are wrong. Consider, for example:

? {1, 2, 3, 4 } * 2  -- prints {2,4,6,8} 

I think the weakness is with your code. I gave you a version that would work with the quantity as an atom. And above, I gave a very simple example of multiplying a sequence and an atom.

Perhaps you should give us an example that doesn't work. Because it does work in general.

Matt

12. Re: Euphoria 4.0.1

• Posted by mattlewis (admin) Mar 13, 2011
• 2590 views

And I responded with a version of the function that works like you seem to want. As I said, the problem isn't with euphoria (as far as I can tell), but with your code. If you really think that euphoria can't multiple (or add or whatever) a sequence with an atom, please give us an example that does not. You did not do that.

Your function fails elsewhere, but not at the point of multiplication. Please show me a simple example, like I did, that gives you an error when doing some arithmetic:

-- test.ex 
 ? {{1,2}, 3, {4.5, 8.6}} * 3 
<eucode> 
 
{{{ 
$ eui test.ex 
{ 
  {3,6}, 
  9, 
  {13.5,25.8} 
} 
}}} 
 
Matt 

14. Re: Euphoria 4.0.1

• Posted by jaygade Mar 13, 2011
• 2608 views

Test results summary: 
Files (run: 334) (failed: 0) (100% success) 
cd ../tests && sh check_diffs.sh 
Both std/memory.e and std/safe.e provide the same interface to std/machine.e....good. 
The include std/safe.e provides as many symbols as as std/memory.e...good. 

I'm running on Linux too, Ubuntu 10.04 32-bit kernel 2.6-32-29-generic. Here's the output, the relevant part I believe:

make test TESTFILE="t_eutest.e" 
cd ../tests && EUDIR=/home/jason/my-euphoria-working-4.0 EUCOMPILEDIR=/home/jason/my-euphoria-working-4.0 \ 
		/home/jason/eubins/eui -i ../include ../source/eutest.ex -i ../include -cc gcc  \ 
		-exe "/home/jason/my-euphoria-working-4.0/source/build/eui" \ 
		-ec "/home/jason/my-euphoria-working-4.0/source/build/euc" \ 
		-eubind "/home/jason/my-euphoria-working-4.0/source/build/eubind" -eub /home/jason/my-euphoria-working-4.0/source/build/eub \ 
		-lib "/home/jason/my-euphoria-working-4.0/source/build/eu.a" \ 
		-verbose t_eutest.e 
interpreting t_eutest.e: 
CMD '/home/jason/my-euphoria-working-4.0/source/build/eui  -i ../include  -d UNITTEST -batch t_eutest.e ' 
interpreting t_c_no_err_file.e: 
sh: eui: not found 
FAILURE: t_c_no_err_file.e No valid control file was supplied. 
 
Test results summary: 
    FAIL: t_c_no_err_file.e 
Files (run: 1) (failed: 1) (0% success) 
interpreting t_c_good_err_file.e: 
sh: eui: not found 
FAILURE: t_c_good_err_file.e No valid ex.err has been generated. 
 
Test results summary: 
    FAIL: t_c_good_err_file.e 
Files (run: 1) (failed: 1) (0% success) 
  failed: eutest t_c_good_err_file.e, expected: 1 but got: 0 
interpreting t_test_true.e: 
sh: eui: not found 
FAILURE: t_test_true.e program died with status 32512 
 
Test results summary: 
    FAIL: t_test_true.e 
Files (run: 1) (failed: 1) (0% success) 
  failed: eutest t_test_true.e, expected: 1 but got: 0 
t_eutest summary: 
  3 tests run, 1 passed, 2 failed, 33% success 
FAILURE: t_eutest.e program died with status 256 

For some reason there's a reference to my eubins eui which I used to ./configure. I'll replace it and go from there. "eubins" is just a link to a current eubin directory I just moved the link to my freshly-made build directory.

Nope, that's not it. It looks like "t_c_no_err_file.e" does not exist. I don't see that file in the default branch, either.

16. Re: Euphoria 4.0.1

• Posted by Vinoba Mar 13, 2011
• 2582 views

I am only trying to point out the fallacy of saying that the simple math functions (+ - * / ) work on sequence against sequence AND sequence against a single element. It does work in an immediate mode as Matt has tried to demonstrate, but hen you have a moderately complicated business need such that you and I have to design a function for an end user, we run into a problem as we have to access each element of a sequence separately and at time the sequence may be a sequence and at times the "sequence" may only be a single value.

Look at the function I wrote above and now we are going to expand it to a moderately relaistic level

500 items in a store. Each item has a separate price Some have discounts (on varying days) Whole section of items (e.g. all tiles) have one single discount as a special which will last only till Friday and then not. Customers come in and buy - sometime 5 items and sometimes 100 items and in small and large quantities. Whilst the cashier is busy, the manger comes to her and says "give him 10% discount on all lumber and 5% on all trusses", the demure old lady next in queue is waiting to buy one screwdriver only with its special discount today. What I have described is a normal store. Yes they all have nicely programmed cash machines so you and I do have to write a program for them, but supposing you were writing an application?

Somewhere in this situation we are faced with having to operate on single values in a sequence and a function with parameters defined as (sequence a, sequence b, sequence c, sequence x, sequence y, sequence z ) does not work if we have to build a loop around each sequence items, because here, there and everywhere, now and then and tomorrow, the sequence supplied will not be sequence but a single element.

Similar problem would arise in a 80 employee office getting payroll. some have exact percentage deductions, some have special deductions, some have holiday pay etc. To accommodate all this in a FUNCTION, you have to work in a loop of a sequence of names or numbers and apply sequences of pay/hour etc. You end up with a loop and your parameter definitions (sequence a,sequence b,sequence c,sequence x,sequence y, sequence z, ) bombs.

18. Re: Euphoria 4.0.1

• Posted by mattlewis (admin) Mar 13, 2011
• 2582 views

And I am afraid you are failing spectacularly. You quoted my working example. Please explain what you think was wrong with it. What do you think it should do?

It's hard to understand what you really want to do here (even taking into account the cashier use case). I can't see what this has to do with sequence vs atom arithmetic.

There's a lot here. There are interface issues. You're implying a lot of filtering (i.e., Lumber vs trusses vs other stuff). This is a lot more complex than the function that calculates prices.

It sounds like you want some magical language construct that knows what you want to do. I can't see any fallacy here, other than that you seem to not understand how euphoria handles sequence vs atom arithmetic. It's not some magical construct. It is either analogous to multiplying a vector by a scalar (sequence multiplied by an atom) or multiplying each element of a sequence with the respective element of another sequence (sequence multiplied by another sequence).

So, once again. Please give us an example of what you think euphoria should do, vs what it actually does. All you've done here is to present a fairly complex use case, and complain that euphoria can't read your mind, or something (sorry, but that's how it's coming across to me.)

I've given you an example that does exactly what you asked about: multiply a sequence by an atom. We can certainly discuss how to write a function using common euphorian idioms, but you're complaining about a basic operation that euphoria has always done, and you're saying it doesn't do it. Even when I've demonstrated that it does. And I also modified your function to do what I thought you were asking about.

Matt

20. Re: Euphoria 4.0.1 (t_eutest.e)

• Posted by jaygade Mar 13, 2011
• 2599 views
• Last edited Mar 14, 2011

Okay, I think that worked. I got 100% success.

Edit: Yes, you're right that I don't have a Euphoria interpreter installed in my path. I removed it while testing the previous build bug and have not reinstalled.

Right now I am making sure that I 'alias' eui to the eui which I want to use and make sure and put a eu.cfg into the working directory.

22. Re: Euphoria 4.0.1

• Posted by evanmars Mar 14, 2011
• 2532 views

Or perhaps the code calling the function could "cast" the ATOM to a SEQUENCE, as in Matt's example, i.e.

if atom( sprice ) then  
	sprice = {sprice)  
end if  

24. Re: Euphoria 4.0.1

• Posted by mattlewis (admin) Mar 14, 2011
• 2490 views

Yes, algorithms often require checking to see whether you have a sequence or an atom. This is inherent in the nature of euphoria objects. If you look at published euphoria code, you'll find many examples. In a previous post, I used the standard library function sum() as a way to abstract this very issue:

public function sum(object a) 
	atom b 
	if atom(a) then 
		return a 
	end if 
	b = 0 
	for i = 1 to length(a) do 
		if atom(a[i]) then 
			b += a[i] 
		else 
			b += sum(a[i]) 
		end if 
	end for 
	return b 
end function 

Sorry, I'm really not trying to beat a dead horse, but claims like "euphoria doesn't do what is says it does" need to be substantiated or proven false (or opened as a ticket, so we can fix a bug!). There's a big difference between that sort of statement and, "That's an interesting feature, but it's not really very useful to me in this context."

Matt

26. Re: Euphoria 4.0.1

• Posted by petelomax Mar 14, 2011
• 2437 views

Do tell, I'm intrigued.

Pete