Menu
▾
▴
#158 Mac OS r6676 CSL reduce.app does not work
Macintosh r6676 CSL reduce.app does not work any more: running "./trunk/bin/redcsl" does not open any window. Running, instead, the command line version "./trunk/bin/redcsl -w" works fine.
This happens both on x86_64 macOS Big Sur v11.7.10 and on M1 macOS Sonoma v14.2.1
The native reduce.app was build with the command
./config --with-csl --disable-universal; make
while the universal reduce.app was build with the command
./config --with-csl --enable-universal; make
Macintosh r6675 CSL reduce.app works well.
Marco
Related
I have checked in continuation of work I was doing to have a BIT more
support for Unicode chars with (CSL) Reduce, and I just rebuild a
universal version on a macbook air m1, so I hope that the errors from
yesterday are sorted. But please check again. Note it would help me if
when you launch the GUI you go "edit/reset font" to get that setting in
line with my changed default for that. Also there a bunch of changes in
character handling so if you see oddities please report. On other
platforms if you enter say "alphaALT-x" you get a greek letter alpha and
further ALT-x cycles through options. On the Mac the ALT key does not seem
to end up detected, but you can use "alpha/ESC/x" instead. Given that that
is not too bad I am not going to dive into ALT key detection right now.
Arthur
On Wed, 3 Jan 2024, Marco Ferraris wrote:
Related
Bugs:
#158Another observation so that a comment is in the mailing list archive. For
a brief time the macports port of brotli stopped providing the static
library I was using, and now that is fixed that static library has a
different name and will not be available unless you have gone
sudo port install brotli-static {+universal if you are using that}
The latest versions of the CSL-reduce configure scripts try to cope with
each situation and thanks are due to those who coordinated with macports
for us on this, but imaginably this could cause issue for some if they get
caught in the transition. Arthur
OK, now both native and universal Macintosh r6679 CSL reduce.app work well. This happens both on x86_64 (2.8 GHz Intel Core i7 quad-core) macOS Big Sur v11.7.10 and on M1 macOS Sonoma v14.2.1.
I compiled also the (1.8 GHz Intel Core i7 quad-core) Cygwin64 Windows 10 Home 64 version of it, which works as expected. Compilation is very slow, probably because of my computer. Once it was faster.
Greek letter are concerned, they work as expected.
It seems that also the menu sub-items "Font…" "Reset Font" work as expected. The default font is DejaVu Sans Mono [PfEd]. I tried to change font to Courier New [TMC] and other monospaced fonts; then I reset the font to default.
Clicking on the CSL Reduce icon in the Dock does not bring the reduce window in front of the other windows. To get the result its enough to click on the XQuartz icon in the Dock. Clicking on a half covered reduce window brings it in front.
After your modification of the code of the procedure 'newtok' in 'newtok.red', and commenting out the two lines of the file 'newtok1.red', it was possible to compile sbcl, clisp and ccl versions of Common Lisp Reduce. It seems that everything is OK when
newtok '((#bullet;) dotprod);
and
newtok '((#times;) crossprod);
are found inside the file 'packages/excalc/vecanlys.red', but not when they are found inside the file 'packages/rlisp/newtok1.red'.
Now I'll see what happens with r6680.
I installed "brotli-static[+unuversal]" as soon as Rainer Müller provided it. I'll hide the new static libraries and let you know what happens.
Marco
Related
Bugs:
#158All your careful testing is very much appreciated.
Maybe they depend on something that is defined after newtok1 is read but
before excalc. newtok1 is the very last file processed within the rlisp
module. If you can coax common lisp into producing a diagnostic and
backtrace that shows what the problem is it can presumably be fixed!
Actually I will wonder if the issue is in rtools and perhaps that
charname.red has not been loaded! If so the fact that the CSL/PSL builds
survive is a bit odd... But there may be other utility functions in the
rtools stuff...
ERROR! The markdown supplied could not be parsed correctly. Did you forget to surround a code snippet with "~~~~"?
Hi Arthur, I’m attaching the relevant parts of the files bootstrat.blg which show the error that the four compilers (abcl, ccl, clisp, sbcl) while compiling the reduce modules. All of them complain about the fact that ‘bullet’ is not, or cannot be coerced to, a character. Marco ======================================================================================================== ABCL +++ Reading file: $reduce/packages/rlisp/newtok1.red nil Error loading ~/sw/reduce-algebra/work/6680/common-lisp/bootstrap.lisp at line 130 (offset 4085) #<THREAD "interpreter" native {4F5CB382}>: Debugger invoked on condition of type TYPE-ERROR The value STANDARD-LISP::BULLET is not of type CHARACTER-DESIGNATOR. [1] SL(1): ======================================================================================================== CCL +++ Reading file: $reduce/packages/rlisp/newtok1.red nil > Error: The value BULLET is not of the expected type (COMMON-LISP:SATISFIES CCL::CHARACTER-DESIGNATOR-P). > While executing: COMMON-LISP:CHARACTER, in process listener(1). > Type :GO to continue, :POP to abort, :R for a list of available restarts. > If continued: Skip loading "bootstrap" > Type :? for other options. ======================================================================================================== CLISP +++ Reading file: $reduce/packages/rlisp/newtok1.red nil *** - COMMON-LISP:CHARACTER: cannot coerce BULLET to a character ======================================================================================================== SBCL +++ Reading file: $reduce/packages/rlisp/newtok1.red While evaluating the form starting at line 129, column 0 of #P"~/sw/reduce-algebra/work/6680/common-lisp/bootstrap.lisp":nil debugger invoked on a COMMON-LISP:SIMPLE-TYPE-ERROR in thread #<THREAD "main thread" RUNNING {10048001D3}>: Symbol name is not of length one: BULLET ======================================================================================================== > Il giorno 5 gen 2024, alle ore 21:50, Arthur Norman <arthurcnorman@users.sourceforge.net> ha scritto: > > All your careful testing is very much appreciated. > > After your modification of the code of the procedure 'newtok' in > 'newtok.red', and commenting out the two lines of the file 'newtok1.red', > it was possible to compile sbcl, clisp and ccl versions of Common Lisp > Reduce. It seems that everything is OK when > newtok '((#bullet;) dotprod); > and > newtok '((#times;) crossprod); > are found inside the file 'packages/excalc/vecanlys.red', but not when > they are found inside the file 'packages/rlisp/newtok1.red'. > > Maybe they depend on something that is defined after newtok1 is read but > before excalc. newtok1 is the very last file processed within the rlisp > module. If you can coax common lisp into producing a diagnostic and > backtrace that shows what the problem is it can presumably be fixed! > > Actually I will wonder if the issue is in rtools and perhaps that > charname.red has not been loaded! If so the fact that the CSL/PSL builds > survive is a bit odd... But there may be other utility functions in the > rtools stuff... > > Now I'll see what happens with r6680. > > I installed "brotli-static[+unuversal]" as soon as Rainer Müller provided it. I'll hide the new static libraries and let you know what happens. > > Marco > > [bugs:#158] Mac OS r6676 CSL reduce.app does not work > > Status: open > Group: > Created: Wed Jan 03, 2024 11:19 PM UTC by Marco Ferraris > Last Updated: Wed Jan 03, 2024 11:19 PM UTC > Owner: nobody > > Macintosh r6676 CSL reduce.app does not work any more: running "./trunk/bin/redcsl" does not open any window. Running, instead, the command line version "./trunk/bin/redcsl -w" works fine. > This happens both on x86_64 macOS Big Sur v11.7.10 and on M1 macOS Sonoma v14.2.1 > > The native reduce.app was build with the command > ./config --with-csl --disable-universal; make > while the universal reduce.app was build with the command > ./config --with-csl --enable-universal; make > Macintosh r6675 CSL reduce.app works well. > > Marco > > Sent from sourceforge.net because you indicated interest in https://sourceforge.net/p/reduce-algebra/bugs/158/ > > To unsubscribe from further messages, please visit https://sourceforge.net/auth/subscriptions/ >Duplicate of #157