Re: [Bug-apl] ⎕FIO[49]

[Juergen Sauermann]

Hi

currently we have 'readlines' in the wrapper function FIO which is contained in the FILE_IO workspace.

BTW the FILE_IO workspace, already contains this wrapper function Z←L FIO[X] R, kindly provided by
Christian Robert, which calls ⎕FIO with string X instead of numeric axis argument X.

Of course I forgot to add FIO∆readlines to the FILE_IO workspace.

/// Jürgen

On 01/27/2017 03:03 AM, Elias Mårtenson wrote:

Thank you Jürgen. You make a compelling argument.

In particular, I'd argue that the ⎕FIO[n] functions should be seen as an internal implementation detail, and the “official” API as it were, should be the wrapper functions.

Speaking of which, ⎕FIO[49] needs a name in wslib5. Any suggestions?

Regards,

Elias

On 25 January 2017 at 19:57, Juergen Sauermann <address@hidden> wrote:

Hi,

there is nothing wrong with ⎕ReadLines as such. However,

- if I remember correctly then in the 70s APL was very much loved by programmers that were
not speaking English every day, in particular in France and Canada.

- Every new ⎕ function requires quite some coding. The C-code that implements it has typically only a few lines
because the function itself is normally already implemented somewhere in libstdio. Suppose you already have the
implementation in, say, { CODE }.

Then making { CODE } available as, say, ⎕FIO[N] is currently simply a new case clause:

   case N:   { CODE }

For ⎕ReadLines instead fo ⎕FIO[N] you would need far more changes:

- a new Id for it,
- a new Token for it,
- a new class, say Quad_ReadLines around { CODE }. That also means two more source code
  files and therefore impacts in src/Makefile.am and SVN
- write testcases for it,
- write a )HELP entry
- write documentation for it

I am normally planning 3 days of (part-time) work for one new Quad Function. This would be 3×49 days
for converting ⎕FIO into different ⎕something functions. And after that, you would have a good
argument to ask why, for example ⎕CR or ⎕SQL are using numbers instead of names.

So this is half a year of work on my side compared to  a simple

)COPY 5 FILE_IO

where named wrapper functions for every ⎕FIO[X] already exist (and you can even change the names
if you don't like them).

/// Jürgen

On 01/25/2017 05:54 AM, Elias Mårtenson wrote:

Is there anything inherently wrong with having more quad-commands? What is the reason we have to do ⎕FIO[49] instead of simply ⎕ReadLines or something like that?

It's like we're going out of way to make code unreadable. I know APL is all about terseness, but I think this is going too far. Am I alone in this?

Regards,

Elias

On 25 January 2017 at 12:36, Christian Robert <address@hidden> wrote:

In fact we should make all/most ⎕FIO[] in into parantheses

things like this was working fine before ⎕FIO[49]

      fd← "r" ⎕fio[3] ".bashrc"
      fd
5
      ⊣ ⎕fio[4] fd
DOMAIN ERROR
      ⊣⎕FIO[4]fd
      ^       ^

well, I know now that I have to make it:

      ⊣ (⎕fio[4] fd)

but it's a change into the previous behavior of ⎕FIO[]. It was not trying to grab the left operator of ⎕fio[4]


Xtian.


On 2017-01-20 10:57, Juergen Sauermann wrote:

Hi Xtian,

yes, one thing that I overlooked was that you should put *⎕FIO**[**49] *into parantheses because
otherwise it will grab the function left of it and apply it to each line read from the file and not to
the final result:

*      ⎕FIO[49]'x'  **
** Line 1 Line 2 Line 3 **
*
*      ⊃⎕FIO[49]'x'     ⍝ apply ⊃ to each line**of the file (has no effect)
** Line 1 Line 2 Line 3 **
**
**      ⊃(⎕FIO[49]'x')   ⍝ apply ⊃ to the final result of ⎕FIO[49]**
**Line 1**
**Line 2**
**Line 3**
*
Syntactically *⎕FIO[49]* behaves like a monadic operator with an optional left function argument.
The parentheses prevent it from using the left argument. In my previous example,

*Z←CONVERT¨Z←⎕FIO**[49] 'filename'
*
the *←* had the same effect as the parentheses.*
*

/// Jürgen


On 01/19/2017 11:10 PM, Christian Robert wrote:

Juergen,

Just a note, your example as of yesterday to disclose the result of ⎕FIO[49] to obtain a matrix no longer work as is.


I had to add a tack (⊢) between ⎕FIO[49] and the "⊃" as in LoadSudoku2 below.


  LoadSudoku2←{{⊂9 9 ⍴ ⍵}⍤1 ⊃ ⊢⎕FIO[49] ⍵}
  LoadSudoku3←{{⊂9 9 ⍴ ⍵} ⎕FIO[49] ⍵}

without the tack it would run disclose on each line which is a noop.


      time "a←LoadSudoku2 'Sudoku_dump.txt'"
3.370101159
      time "a←LoadSudoku3 'Sudoku_dump.txt'"
2.977278584

      time "a←LoadSudoku2 'Sudoku_dump.txt'"
3.142475014
      time "a←LoadSudoku3 'Sudoku_dump.txt'"
2.966579181

the LoadSudoku3 run faster than the LoadSudoku2


      ⍴a
219025
      ⊃a[1 2 3]
 000000000  000000000  000000000
 000000008  000000009  000000009
 560230400  400008120  580030100
 100803000  001030000  001640008
 090060000  085009760  000300005
 000025010  090040200  602005041
 300500706  800070006  003700000
 076002000  000021050  920000076
 005090021  006000010  040809000
      Display ¨ ⊃a[1 2 3]
 ╔═══╤═══╤═══╦═══╤═══╤═══╦═══╤═══╤═══╗  ╔═══╤═══╤═══╦═══╤═══╤═══╦═══╤═══╤═══╗  ╔═══╤═══╤═══╦═══╤═══╤═══╦═══╤═══╤═══╗
 ║   │   │   ║   │   │   ║   │   │   ║  ║   │   │   ║   │   │   ║   │   │   ║  ║   │   │   ║   │   │   ║   │   │   ║
 ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢
 ║   │   │   ║   │   │   ║   │   │ 8 ║  ║   │   │   ║   │   │   ║   │   │ 9 ║  ║   │   │   ║   │   │   ║   │   │ 9 ║
 ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢
 ║ 5 │ 6 │   ║ 2 │ 3 │   ║ 4 │   │   ║  ║ 4 │   │   ║   │   │ 8 ║ 1 │ 2 │   ║  ║ 5 │ 8 │   ║   │ 3 │   ║ 1 │   │   ║
 ╠═══╪═══╪═══╬═══╪═══╪═══╬═══╪═══╪═══╣  ╠═══╪═══╪═══╬═══╪═══╪═══╬═══╪═══╪═══╣  ╠═══╪═══╪═══╬═══╪═══╪═══╬═══╪═══╪═══╣
 ║ 1 │   │   ║ 8 │   │ 3 ║   │   │   ║  ║   │   │ 1 ║   │ 3 │   ║   │   │   ║  ║   │   │ 1 ║ 6 │ 4 │   ║   │   │ 8 ║
 ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢
 ║   │ 9 │   ║   │ 6 │   ║   │   │   ║  ║   │ 8 │ 5 ║   │   │ 9 ║ 7 │ 6 │   ║  ║   │   │   ║ 3 │   │   ║   │   │ 5 ║
 ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢
 ║   │   │   ║   │ 2 │ 5 ║   │ 1 │   ║  ║   │ 9 │   ║   │ 4 │   ║ 2 │   │   ║  ║ 6 │   │ 2 ║   │   │ 5 ║   │ 4 │ 1 ║
 ╠═══╪═══╪═══╬═══╪═══╪═══╬═══╪═══╪═══╣  ╠═══╪═══╪═══╬═══╪═══╪═══╬═══╪═══╪═══╣  ╠═══╪═══╪═══╬═══╪═══╪═══╬═══╪═══╪═══╣
 ║ 3 │   │   ║ 5 │   │   ║ 7 │   │ 6 ║  ║ 8 │   │   ║   │ 7 │   ║   │   │ 6 ║  ║   │   │ 3 ║ 7 │   │   ║   │   │   ║
 ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢
 ║   │ 7 │ 6 ║   │   │ 2 ║   │   │   ║  ║   │   │   ║   │ 2 │ 1 ║   │ 5 │   ║  ║ 9 │ 2 │   ║   │   │   ║   │ 7 │ 6 ║
 ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢  ╟───┼───┼───╫───┼───┼───╫───┼───┼───╢
 ║   │   │ 5 ║   │ 9 │   ║   │ 2 │ 1 ║  ║   │   │ 6 ║   │   │   ║   │ 1 │   ║  ║   │ 4 │   ║ 8 │   │ 9 ║   │   │   ║
 ╚═══╧═══╧═══╩═══╧═══╧═══╩═══╧═══╧═══╝  ╚═══╧═══╧═══╩═══╧═══╧═══╩═══╧═══╧═══╝  ╚═══╧═══╧═══╩═══╧═══╧═══╩═══╧═══╧═══╝


I love ⎕FIO[49] ;-)

Xtian.


On 2017-01-19 09:17, Juergen Sauermann wrote:

Hi,

I have added *⎕FIO[49]* to GNU APL in *SVN 858*.

*Z←⎕FIO[49] 'filename'* reads the file named *filename* line by line and stores every line
of the file in one enclosed item of result vector*Z* , for example (file *x* has 3 lines):

*      **⍪**⊃ ⎕FIO[49] 'x'
 Line 1
 Line 2
 Line 3
*
The end of line character (LF = '\n') character possibly a trailing CR = '\r' character is being removed in the process.

*⎕FIO[49]* accepts an optional left function argument *F* which is a monadic conversion function.
The conversion functions F is called with every line read and the result of the function
is then stored in Z. for example:

*      ⍪⊃ {'x: ', ⍵} ⎕FIO[49] 'x'**
**x: Line 1**
**x: Line 2**
**x: Line 3*

/// Jürgen