Raistlin Magere wrote:
Hi, I tried to run Cello through the cloucell.lpr (after having apprpriately modified the hard-links); it complains about the lack of a uffi.lpr file, as I already have uffi installed on my image {(asdf:operate 'asdf:load-op :uffi)} I just removed the link to it in the opengl-ffi.lpr {i.e. I remove the line :projects (list (make-instance 'project-module :name "..\uffi\uffi"))}
Puzzling how that got left out. Current one is attached, but it just builds UFFI the way ASDF would. I am perty sure I still have acl501 on my old NT box. I will fire that up later and see if I can make that port.
However when I then run cloucell I get the list of errors that enclose at the end of this message. As you can see most errors regard the fact that unsigned-integer is not of the right type. There are also some other errors though; mainly Part4 and Part11 about missing files, Part5 about a function requiring exactly 9 parameters and being passed only 8 Part10 about a missing class (though this could be due to previous errors) Part8 and Part9 warnings about free variable reference (not a big deal) Part12 about While --- this error has been in cells for a very long time there is only 1 while which needs to be replaced with either an if call or with a cl::while when running cells in acl5.0.1 The problem is in dataflow-management.lisp and the line is (while (and *sw-pending* (not *sw-looping*)) #+dfdbg
Now that I am a looper I can just use that.
Just as a side note I am using ACL5.0.1 (which could explain problem about Part5 as maybe the functions requirement have changed from 5.0.1 to 6.2) also in order to get Uffi to run in 5.0.1 the line: :strings-convert nil) in def-function from the file "functions.lisp" of the uffi package, had to be removed; has such parameter doesn't exist in 5.0.1 while is present in 6.2 -not sure if this can explain the unsigned-integer errors.
Ah, yes, vague memory on that. Thx for the reminder. btw, did you grab the latest UFFI? Hopefully it is backward compatible to acl501 (assuming it has been updated for acl62.)
Maybe I should try running it again by using your uffi.lpr file (I don't have it though).
Thanks, raistlin
Thank you! So much for the "null port". :) I should have known better.
Here are the errors they are all marked by ---Errors Partx--- where x is the number I refered to in the above message.
---Errors Part1--- ; While compiling GL-MATERIALFV in c:\Cello-Dev\cello\opengl-ffi\gl-functions.lisp: Warning: a direct call of `(glMaterialfv LANGUAGE C)' cannot be compiled in-line because the argument type `UNSIGNED-INTEGER' is not either a one-dimensional simple-array or one of: :FOREIGN-ADDRESS :LISP FIXNUM INTEGER SINGLE-FLOAT DOUBLE-FLOAT :SINGLE-FLOAT-NO-PROTO SIMPLE-STRING CHARACTER.
I am trying to remember if that is necessarily bad, or if I get them myself.
---Errors Part4--- Cannot find c:\Cello-Dev\cello\opengl-ffi\nehe-14.lisp. Will skip and continue compiling project.
I forgot that was pulled in my the bindings project when I was downsizing the zip to get it thru. Anyway, looks like it needs work, it exits and takes down ACL just after opening the window. You guessed right, it is a throwaway, but maybe I should bring it back since it is a nice half-port: it only requires freeglut (or some glut). Anyway, the first thing I will do is get all this sorted out /and/ put the next zip up on tilton-technology with a password (unless I can figure out SSH again and just FTP it up to the common-lisp project. but then I also have to figure out the GnuPG thing because they want me to sign this stuff. :()
---Errors Part5--- ; While compiling CELLS-TIME-REPORT in c:\Cello-Dev\cells\debug.lisp: Warning: EXCL::TIME-REPORT should be given exactly 9 arguments. It was given 8 arguments. Problem detected when processing (EXCL::TIME-REPORT CPU-GC-USER CPU-GC-SYS ...) inside (BLOCK CELLS-TIME-REPORT (TRC "cpu-gc-user" CPU-GC-USER) ...) inside (PROGN (BLOCK CELLS-TIME-REPORT (TRC "cpu-gc-user" CPU-GC-USER) ...))
Omigod! That is totally ACL specific. I have to feature that out with +ACL62 or even just lose it.
---Errors Part6--- ; While compiling MAGICK-CONTRAST-IMAGE in c:\Cello-Dev\cello\magick-wand.lisp: Warning: a direct call of `(MagickContrastImage LANGUAGE C)' cannot be compiled in-line because the argument type `UNSIGNED-INTEGER' is not either a one-dimensional simple-array or one of: :FOREIGN-ADDRESS :LISP FIXNUM INTEGER SINGLE-FLOAT DOUBLE-FLOAT :SINGLE-FLOAT-NO-PROTO SIMPLE-STRING CHARACTER
Looks like this is our only hard error. While playing with nehe-14 just now I did a full build and got no errors on anything.
---Errors Part8--- ; While compiling (:INTERNAL (:INTERNAL MAKE-LIGHTING 1) 0) in c:\Cello-Dev\cello\lighting.lisp: Warning: Free reference to undeclared variable *MGW-ZNEAR* assumed special. Warning: Free reference to undeclared variable *MGW-ZFAR* assumed special.
Looks familiar. I will fix (or determine if it is an ASDF dependency thing).
---Errors Part10--- Error: No class named: EXCL::STRUCTURE-EFFECTIVE-SLOT-DEFINITION. [condition type: PROGRAM-ERROR]
My bad.
---Errors Part11--- Cannot find c:\Cello-Dev\cello\load-music.lisp. Will skip and continue compiling project Cannot find c:\Cello-Dev\cello\visual-music.lisp. Will skip and continue compiling projec Cannot find c:\Cello-Dev\cello\vm-test.lisp. Will skip and continue compiling project.
Oy, just lose those (or better, just wait till next distro this PM). Cello does not play (yet!).
---Errors Part12--- While compiling these undefined functions were referenced: CELLS::WHILE.
I'll nail it.
Thx, sorry for the avoidable errors, new distro coming ASAP (but I am hitting the gym first).
kenny