mingw64 build patch

[xytroxon]

Doing a write-file of the "test" string shows no premature shortening of the string or any extra "odd" chars in the file on disk.

Code Select Expand


(write-file "test.txt" test)


But reading and printing the file has the same display failure.

Code Select Expand


(print (read-file "test.txt"))


So maybe the problem lies within the print function.



--xytroxon

[TedWalther]

I remember a few years ago, Phillip Whelan was able to very quickly locate and patch a 64 bitness problem in newLisp.  Wish he was still around, this sounds like a 64 bitness type of problem, right up his alley.

[Lutz]

Cannot repeat this problem. Your test code or printing any other string >= 6000 characters always works on Windows 7 Home Premium SP1 command shell  cmd.exe or using the MSYS/MinGW bash shell. I wonder if other Windows 7 users can repeat this problem. What version of Windows are you using?

[HPW]

Hello,



I tried it on my win7 home premium tablet.

When I paste it line by line into the newlisp Shell it works as expected.



But when i make a file Test.lsp with:

Code Select Expand


(setq a (string (dup "+" (* 80 25))))
(setq b (string (dup "-" (* 80 25))))
(setq c (string (dup "*" (* 80 25))))
(setq test (string a b c))

(silent(print (length test) "n" test))


and do a (load "Test.lsp") in a fresh started Shell I get:

Code Select Expand


newLISP v.10.6.3 64-bit on Windows IPv4/6 libffi, options: newlisp -h

> (load "Test.lsp")
6000
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
--------------------------------------------------------¦.6Å-ï?


This seems to be Buggy.



Edit: The 32 bit Exe does this correctly.



Regards

Hans-Peter

[Lutz]

Yes, now I can repeat it too when loading the test code from a file. I could not repeat it when testing the code interactively, but it also fails when interactively when using silent.



It also fails on the 32-bit version in 10.6.3. It seems that this has nothing to too with 32-bit versus 64-bit, but is related to other changes in 10.3.6 development. I am glad we are catching this before 10.6.3 goes out of "inprogress".

[kosh]

I also confirmed this IO problem.

Code Select Expand

$ newlisp64.exe -n
> (setf str (string (dup {X} 2049) {HELLO}))
[text]XXXX...<repeat 2049 times>HELLO[/text]
> (println str)
XXXX...<repeat 2049 times>


It seems that cause internal output function `my_vasprintf` (in nl-filesys.c, called from varPrintf)



`vasprintf` has been implemented in mingw64 in stdio.h (but mingw32 yet).

then, It could be fixed in this way:

Code Select Expand

/* newlisp.h */
#ifdef WINDOWS
#ifndef __MINGW64__ /* macros depends on mingw implementation */
# define MY_VASPRINTF
# define vasprintf my_vasprintf
#endif
...
#endif


Regards.

[Lutz]

Yes, it was vasprintf, thanks Kosh:



http://www.newlisp.org/downloads/development/inprogress/">http://www.newlisp.org/downloads/develo ... nprogress/">http://www.newlisp.org/downloads/development/inprogress/

[ValentiMr]

Attached is a patch to fix the test build on head.  All it does it make it so the tests compile:  they still don't pass at least they don't pass on mac os x.



Also, I've noticed a bunch of build issues on Mac OS X.  Anyone mind if I fix those?



