Hi Blake,
thanks, fixed in SVN 371.
/// Jürgen
On 07/12/2014 06:27 PM, Blake McBride wrote:
)CLEAR
CLEAR WS
∇test
[1] 5∇
SYNTAX ERROR
[1]
You should be able to end a line with a del to close the function and
end the edit. It works on IBM APL 2.
Thanks.
Blake
Greetings,
It would be very helpful to have the following C functions added to the
FILE_IO interface:
getcwd - good to be able to tell where you are
access - good to test to see if a file exists
Thanks.
Blake
Greetings,
The preferences file contains name/value pairs. It would be nice to be
able to specify arbitrary name/value pairs that can be accessed from within
APL somehow. For example, let's say you store your component files an a
standard place, or you use a certain default SQL database. It wou
Looks good. Thanks!
On Mon, Jul 14, 2014 at 5:37 AM, Juergen Sauermann <
juergen.sauerm...@t-online.de> wrote:
> Hi Blake,
>
> thanks, fixed in SVN 371.
>
> /// Jürgen
>
>
>
> On 07/12/2014 06:27 PM, Blake McBride wrote:
>
>)CLEAR
> CLEAR WS
> ∇test
> [1] 5∇
> SYNTAX ERROR
> [1]
Greetings,
I installed GNU APL on raw LinuxMint 16 and 17 and came up with the
following differences from what is stated in the README-1-prerequisits
file. Note that since LinuxMint is built on to of Ubuntu, Ubuntu installs
should be the same.
1. Although ./configure worked out-of-the-box witho
/Volumes/Data/Development/MyProjects/GNUAPL/apl-svn/src/APserver.cc:609:1:
Control reaches end of non-void function
#else
int
open_UNIX_socket(const char * listen_name)
{
cerr << "*** this platform does not support address family AF_UNIX!";
}
#endif
Hi Blake,
I believe the ⎕ES related bugs reported recently should be fixed in SVN 372.
This one is not, though. The reason is that fixing it would have
considerable performance
impacts (we would need a copy of every defined function argument only
for the rare case that ⎕ES
is called),
I also
Jürgen,
Do you think it would be possible to implement some way of retrieving not
only the error code, but also the )MORE text from an exception inside
an ⎕EA?
The reason for this is that in my SQL∆WithTransaction call, I need to call
the rollback function if there is an error, but then I want to
Works. Thanks!
On Sun, Jul 13, 2014 at 12:27 PM, Blake McBride wrote:
> GNU APL:
>
> )CLEAR
> CLEAR WS
> ∇test
> [1] '⎕ES''Hello'''⎕EA'3÷0'
> [2] ∇
> test
> Hello
> '⎕ES'Hello'' ⎕EA '3÷0'
> ^^
>
>
> IBM APL 2:
>
> )CLEAR
> CLEAR WS
> ∇TEST
On 2014-07-14 16:15:32, Juergen Sauermann wrote:
> Since ⎕ES is not ISO-standard anyhow, I take the freedom to trade performance
> for APL2 compatibility.
But it is. See sections 11.5.7 and 11.6.5.
And standard says it should return nil for empty right
argument, which means not returning any valu
Fixed. This was a problem because when ]boxing was turned on, I kept
seeing all those lines.
Thanks!
On Sun, Jul 13, 2014 at 1:24 PM, Blake McBride wrote:
> x←⎕ES''
>
> works on GNU APL. It gives a VALUE ERROR in IBM APL 2. ⎕ES should not
> return a value.
>
> Thanks.
>
> Blake
>
>
That "return nil" is clearly a typo and a further reflection of the poor
quality of that spec. I offer the following as proof that that comment is
a typo:
1. Look at the signature of ⎕ES at the beginning of 11.5.7. It says:
⎕ES B
it does not say:
Z←⎕ES B
like they do everywhere el
Dear Juergen,
Every place that I use ⎕ES I am being forced to treat the arguments as
read-only. I also had to re-order code to account for optional arguments.
It makes for some ugly code. I hope when you add lazy copying of arrays,
the internal code to do this correctly won't be a problem.
Show
Hi,
thanks, fixed in SVN 373.
/// Jürgen
On 07/14/2014 05:36 AM, Elias Mårtenson wrote:
Interesting. Given the following definition of pp:
∇Z←X pp Y
⎕←'comparing'
⎕←' X=' (8⎕CR X)
⎕←' Y=' (8⎕CR Y)
Z←X≡Y
∇
I get the following output:
*(⊂'foo') pp¨ (,⊂'foo')*
compar
Hi Blake,
thanks, added in SVN 373. Left argument of access() is a string
consisting of
R, W, X, and F rather than a single integer.
/// Jürgen
On 07/14/2014 01:21 PM, Blake McBride wrote:
Greetings,
It would be very helpful to have the following C functions added to
the FILE_IO interface
Hi Peter,
thanks, should be fixed in SVN 373.
/// Jürgen
On 07/14/2014 03:34 PM, Peter Teeson wrote:
/Volumes/Data/Development/MyProjects/GNUAPL/apl-svn/src/APserver.cc:609:1:
Control reaches end of non-void function
#else
int
open_UNIX_socket(const char * listen_name)
{
cerr << "*** t
You can use popen() to run pwd in a shell. It'll give you the same answer
as calling getcwd() from within the APL process.
A file's presence can be probed using fstat(). A negative return value
denotes the absence of the file.
popen() and fstat() are already exposed via lib_file_io.
Take a look
Whoops, this thread was about the message rather than return value.
About that, standard states that event-message is local to a context and
that ⎕ES should create an exception in caller's context, but I don't
think it's properly specified what does it exactly mean.
But the point is, that ⎕ES _is_
Hi,
should be fixed in SVN 374.
The sqlite should be optional - Elias, there is a
*#include "apl-sqlite.hh"*
in *ResultValue**.hh* - that looks wrong.
*execinfo* is a separate lib on some platforms and part of *glibc* on
others.
It is OK if the test for it fails.
/// Jürgen
iOn 07/14/201
Got it. Thanks.
On Mon, Jul 14, 2014 at 11:44 AM, Kacper Gutowski
wrote:
> Whoops, this thread was about the message rather than return value.
> About that, standard states that event-message is local to a context and
> that ⎕ES should create an exception in caller's context, but I don't
> thi
Elias' thread about docstrings got me to thinking about other function
metadata.
One thing that might be nice to have is for APL to record the source
location of a function's definition.
If the function is defined in a file, record the path and the line number
of the first line of the definition.
Hi,
Kacper is right - it is in the standard.
But the standard merely says "Set the rest of event-message as appropriate".
The rest is rather APL2 specific. I could also argue that the current
value of the parameter
is more relevant to the exception than the original value. I have put
this int
Hi,
my favourite for code documentation is *Doxygen*. It does not currently
support APL comments but we might be able to change that. Or use one
of the existing tags like --- for *VHDL* or %%% for *Erlang*.
It would need some changes in GNU APL - multi-line comments and
proper storage of documen
A component file system for GNU APL has been released for GNU APL. It is
located at:
https://github.com/blakemcbride?tab=repositories
This Component File System fully conforms to:
Programming Language APL, Extended
ISO/IEC 13751:2000(E), Annex A
It is built on top of the SQL Interface by Elias
You must set errno = 0 before calling access(). Otherwise the lib_file_io
return value may be incorrect upon success.
-- Forwarded message --
> From: Juergen Sauermann
> To: Blake McBride
> Cc: "bug-apl@gnu.org"
> Date: Mon, 14 Jul 2014 18:01:09 +0200
> Subject: Re: [Bug-apl] A
I apologize for the confusion, but this proposal is about capturing the
file location of a function's definition; not about the function's
comments. My lead sentence was just an indicator that the earlier
discussion of comments triggered this new line of thought...
On Mon, Jul 14, 2014 at 11:14 A
libncurses-dev should be libncurses5-dev
On Mon, Jul 14, 2014 at 12:36 PM, Juergen Sauermann <
juergen.sauerm...@t-online.de> wrote:
> Hi,
>
> should be fixed in SVN 374.
>
> The sqlite should be optional - Elias, there is a
>
> *#include "apl-sqlite.hh"*
>
> in *ResultValue**.hh* - that looks
This information is already stored. The Emacs mode uses this when
navigating to definition of a function.
Regards,
Elias
On 15 Jul 2014 02:19, "David Lamkins" wrote:
> I apologize for the confusion, but this proposal is about capturing the
> file location of a function's definition; not about th
That's good to know. Thanks!
On Mon, Jul 14, 2014 at 4:34 PM, Elias Mårtenson wrote:
> This information is already stored. The Emacs mode uses this when
> navigating to definition of a function.
>
> Regards,
> Elias
> On 15 Jul 2014 02:19, "David Lamkins" wrote:
>
>> I apologize for the confus
Did you mean ? Then that is actually correct. ResultValue is
specific to the SQLite provider. If SQLite support is not enabled, that
file shouldn't be compiled.
Perhaps it would be best if that file is renamed SqliteResultValue...
Regards,
Elias
On 15 July 2014 01:36, Juergen Sauermann
wrote:
The thing that started this is that GNU APL was not able to compile without
the SQLite .h files. In other words, GNU APL requires SQLite. I think he
wants it to only use SQLite when ./configure finds it.
On Mon, Jul 14, 2014 at 9:02 PM, Elias Mårtenson wrote:
> Did you mean ? Then that is act
It's not supposed to. I think the error is that it still tries to compile
ResultValue.cc even though SQLite wasn't found during the ./configure phase.
It's most definitely supposed to (and it was written to) not require either
SQLite *nor* Postgres to build.
Regards,
Elias
On 15 July 2014 10:09
Jürgen,
I have updated the SQL module in the git repository. First and foremost I
have renamed ResultValue to SqliteResultValue to better illustrate that
it's only required by the SQLite provider. Secondly, I've added
documentation to the sql.apl library.
This integrates neatly with the new docum
Just some extra information: There is not currently (at least as far as I
know) any way you can read this information from within APL. You can,
however, set it using the dyadic form of ⎕FX:
*0 'foo.c:10' ⎕FX 'Z←aa X' 'Z←X+1'*
This specifies that the function aa was defined on line 10 of the f
And of course, defining a function using *C-c C-c*
(gnu-apl-interactive-send-current-function) defines the function using the
dyadic ⎕FX as described above. The )LOAD function in GNU APL also sets it.
And there's how function navigation is implemented in the Emacs mode. :-)
Regards,
Elias
On 15
35 matches
Mail list logo