openj/core

abandoned, please check out https://github.com/jsoftware/jsource

core's Issues

plot fails on windows for user folders with spaces in name

   require'plot'
   plot 3 4 5
view pdf error:c:/program files/ghostgum/gsview/gsview64.exe c:/users/conor o'brien/j64-804-user/temp/plot.pdf

The system cannot find the file specified.

The pdf saves fine, I just suspect that it's because my root name is conor o'brien. Please fix this ASAP.

Android version of J has intermittent scrolling issue.

The video I uploaded here pretty much shows what I'm running into, where the transcript "snaps" back to the cursor without giving me a chance to select appropriate text for editing.

https://www.youtube.com/watch?v=hCdw7MHj-2k

launchpad

openj is a project on launchpad:
https://launchpad.net/openj
at the least, we can send patches there.
might be friendlier for some people.

10 vote for autotools and /no/ CMake

there has been discussions about build systems
for openj. also, there's discussion on the Jsource
mailing list. please vote up the issue you choose.

if you're new to github (like me) you can vote
(and retract a vote) by clicking on the up-arrow
to the left of the issue.

i've created 4 issues:
* 00 vote for /neither/ autotools nor CMake
* 01 vote for CMake and /no/ autotools
* 10 vote for autotools and /no/ CMake
* 11 vote for both autotools /and/ CMake

please note that this is a discussion about
including a build system in core. we can
manage other build systems by creating
repos for them in openj or by branching
openj/core (it would be a small divergence)

failing tests

This is no good.

One problem is the lack of a make test. In fact, the lack of a working jconsole without having to make install is a problem, too. In future, openj requires a make test report for at least linux and windows -- ideally, we borrow a build-and-test farm from somewhere to automate the checking of all pull requests on all platforms.

Nothing but bug-fixes or reverts until we're clean.

make: work on the build system

we could write a unix Makefile, and an nmake Makefile quite easily using J.
this way, we won't have the double maintenance of two slightly different files.
i would favour this approach this if there is no need for VC++ projects/solution.
however, i reckon there is that need.

i'm not brave enough to try it, but maybe someone would like to see if it's
feasible to generate VC++ projects/solution in the same way as generating gmake
or nmake makefiles.

11 vote for both autotools /and/ CMake

there has been discussions about build systems
for openj. also, there's discussion on the Jsource
mailing list. please vote up the issue you choose.

if you're new to github (like me) you can vote
(and retract a vote) by clicking on the up-arrow
to the left of the issue.

i've created 4 issues:
* 00 vote for /neither/ autotools nor CMake
* 01 vote for CMake and /no/ autotools
* 10 vote for autotools and /no/ CMake
* 11 vote for both autotools /and/ CMake

please note that this is a discussion about
including a build system in core. we can
manage other build systems by creating
repos for them in openj or by branching
openj/core (it would be a small divergence)

separate 'core' j from the 'bells and whistles'

this may be misguided...

test for core team notifications

test

missing newlines

need to do a simple
$ find . -exec dos2unix {} ;

to add terminating newlines to files that don't have them.

01 vote for CMake and /no/ autotools

there has been discussions about build systems
for openj. also, there's discussion on the Jsource
mailing list. please vote up the issue you choose.

if you're new to github (like me) you can vote
(and retract a vote) by clicking on the up-arrow
to the left of the issue.

i've created 4 issues:
* 00 vote for /neither/ autotools nor CMake
* 01 vote for CMake and /no/ autotools
* 10 vote for autotools and /no/ CMake
* 11 vote for both autotools /and/ CMake

please note that this is a discussion about
including a build system in core. we can
manage other build systems by creating
repos for them in openj or by branching
openj/core (it would be a small divergence)

configure should handle --prefix

You should be able to specify a destination to make install to. Configure with --prefix is the first part of this.

Android version of J crashes repeatedly with faulty code.

While working with the Android version of J701, referenced from http://www.jsoftware.com/jwiki/Android, I've discovered certain classes of syntax errors will cause the interpreter to crash hard. I downloaded the APK here: https://github.com/mdykman/jconsole_for_android/blob/master/dist/j-console.apk

I would have left this issue on mdykman's Github page, but no issues tab exists for it.

To reproduce:

a=:a
a''

When executed, the above code will result in the J interpreter force-stopping and returning back to the Android home screen. I would rather expect it to deliver a "syntax error" instead, as it does on the desktop.

I know it's not an infinite loop exceeding its stack, because this:

a=:$:
a''

will properly deliver a stack error.

Thank you!

source distributions

i'd like to get some source distributions out to the Jsource community. i need to re-read the Jsource archive to make sure i catch everything they've been talking about with getting windows compiles to work.

last time i checked cmake, you needed cmake installed to build the VC++ projects/solution that cmake produces. is it possible to send out source for VC++ without the end user needing to install cmake?

Build on Android

I'm going to take a stab at building on Android. I have literally no idea what this entails or even if it's possible, but I'm Sure It Will Be Fine, Right?

Can't compile J with tcc

I tried to compile J with the tiny C compiler (tcc). Compilation failed. Here is a detailed description of all errors:

http://www.mail-archive.com/[email protected]/msg03685.html

Build failure on gcc 5.2.0

• Linux 4.3.0 x86_64
• gcc 5.2.0

Attempting to build results in the following error:

/home/earnest/build/lab/openj-git/src/core/f2.c: In function ‘jtfmt1’:
/home/earnest/build/lab/openj-git/src/core/f2.c:115:4: error: non-floating-point argument in call to function ‘__builtin_isnan’
    else if(_isnan(*wv)          )strcpy(jt->th2buf,e?"  _.":' '==*s?" _.":"_.");
    ^
CMakeFiles/j.dir/build.make:926: recipe for target 'CMakeFiles/j.dir/f2.c.o' failed
make[2]: *** [CMakeFiles/j.dir/f2.c.o] Error 1
make[2]: *** Waiting for unfinished jobs....
/home/earnest/build/lab/openj-git/src/core/i.c: In function ‘jtmakename’:
/home/earnest/build/lab/openj-git/src/core/i.c:25:11: warning: implicit declaration of function ‘nmhash’ [-Wimplicit-function-declaration]
  zv->hash=nmhash(m,s); 
           ^
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/j.dir/all' failed
make[1]: *** [CMakeFiles/j.dir/all] Error 2
Makefile:127: recipe for target 'all' failed
make: *** [all] Error 2

00 vote for /neither/ autotools nor CMake

there has been discussions about build systems
for openj. also, there's discussion on the Jsource
mailing list. please vote up the issue you choose.

if you're new to github (like me) you can vote
(and retract a vote) by clicking on the up-arrow
to the left of the issue.

i've created 4 issues:
* 00 vote for /neither/ autotools nor CMake
* 01 vote for CMake and /no/ autotools
* 10 vote for autotools and /no/ CMake
* 11 vote for both autotools /and/ CMake

please note that this is a discussion about
including a build system in core. we can
manage other build systems by creating
repos for them in openj or by branching
openj/core (it would be a small divergence)

voting testing

can i vote on this??