Re: Maxima works with gcl 2.7.1

I sent the make.log showing the gcl_typep.lsp crash yesterday, here it is
again: https://pastebin.com/EqkjDwaF
This is the crash where I used GDB as you requested and it showed an
endless recursion:

#0  0x00000000000a7e69 in maybe_reallocate_page ()
#1  0x00000000000afbc5 in alloc_object ()
#2  0x0000000000098ba8 in alloc_simple_string ()
#3  0x0000000000098f21 in make_simple_string ()
#4  0x00000000000afc55 in alloc_object ()
#5  0x0000000000098ba8 in alloc_simple_string ()
#6  0x0000000000098f21 in make_simple_string ()
#7  0x00000000000afc55 in alloc_object ()
#8  0x0000000000098ba8 in alloc_simple_string ()
#9  0x0000000000098f21 in make_simple_string ()
#10 0x00000000000afc55 in alloc_object ()
[repeating for a very long time]

Unfortunately, I can't do unixport/saved_gcl3 right now, since I don't
manage to get far enough with the build any more ... Maybe it's just luck
and I have to keep trying.

Am Fr., 25. Apr. 2025 um 23:11 Uhr schrieb Camm Maguire <
[email protected]>:

> Greetings!
>
> Clearly something non-reproducible is going on.
>
> Please do
>
> unixport/saved_gcl3
> >(compile-file "lsp/gcl_s.lsp")
> >(si::bye)
> ls -lrt lsp/gcl_s.*
>
> If you do not see .lsp, .o, .c, .h, and .data files, then
>
> unixport/saved_gcl3
> >compiler::*default-c-file*
> >compiler::*default-h-file*
> >compiler::*default-data-file*
> >compiler::*default-system-p*
>
> I don't know what to say about the crash at gcl_typep.lsp, as I don't
> think I've seen a log showing this (yet).
>
> Take care,
>
> David Scherfgen via Maxima-discuss
> <[email protected]> writes:
>
> > Yes indeed, unixport/saved_ansi_gcl seems to be working!
> >
> > $ ls -lrt gcl/gcl_s.*
> > -rw-rw-r-- 1 gcl-builder gcl-builder 61813 Apr 25 08:32 gcl/gcl_s.o
> > -rw-rw-r-- 1 gcl-builder gcl-builder 43685 Apr 25 08:32 gcl/gcl_s.data
> >
> > $ ls -lrt gcl3/gcl_s.*
> > -rw-rw-r-- 1 gcl-builder gcl-builder 43685 Apr 25 08:32 gcl3/gcl_s.data
> > -rw-rw-r-- 1 gcl-builder gcl-builder 61813 Apr 25 08:32 gcl3/gcl_s.o
> >
> > Regarding the missing file, a hardware issue seems unlikely, as I'm
> using ZFS on a 2x NVMe mirror.
> >
> > Also strange: I just tried to build again with exactly the same setup as
> before, and this time it crashed while compiling lap/gcl_typep.lsp, just
> like yesterday.
> >
> > Another attempt, after rebooting, resulted in the same problem.
> >
> > Camm Maguire <[email protected]> schrieb am Fr., 25. Apr. 2025,
> 22:01:
> >
> >  Greetings!  Your build basically finished but for the final profiling
> >  images.  You likely have a working unixport/saved_ansi_gcl.
> >
> >  The error seems to point to some os or hardware failure
> >  writing/retaining a file.  Please send `ls -lrt gcl/gcl_s.*` and `ls
> >  -lrt gcl3/gcl_s.*`.  The .o file is there, but you seem to be missing
> >  the .c file.  And your log shows 'cp gcl3/gcl_s.* gcl' succeeding.
> >
> >  Take care,
> >
> >  David Scherfgen via Maxima-discuss
> >  <[email protected]> writes:
> >
> >  > Oh, now it fails at a different point.
> >  > config.log: https://pastebin.com/vvLeHVaK
> >  > make.log: https://pastebin.eldshort.de/bQyxGMAP_4YlYA1-hhzyK
> (exceeds pastebin.com's size limit)
> >  > The error is now: "cc1: fatal error: gcl/gcl_s.c: No such file or
> directory"
> >  >
> >  > Camm Maguire <[email protected]> schrieb am Fr., 25. Apr. 2025,
> 20:11:
> >  >
> >  >  Greetings!  I believe your previous attempt failed earlier than this.
> >  >  Can you please refresh the pastebins?
> >  >
> >  >  Take care,
> >  >
> >  >  David Scherfgen via Maxima-discuss
> >  >  <[email protected]> writes:
> >  >
> >  >  > Unfortunately, configure --enable-min-pagewidth=21 didn't help. It
> still crashes the same way while compiling gcl_typep.lsp. Of course I made
> sure to
> >  >  clean up
> >  >  > everything before configuring and making.
> >  >  >
> >  >  > Camm Maguire <[email protected]> schrieb am Fr., 25. Apr.
> 2025, 04:38:
> >  >  >
> >  >  >  Greetings!  OK, I've just tested this, and thankfully you can
> just use
> >  >  >  2.7.1 with a configure option:
> >  >  >
> >  >  >  ./configure --enable-min-pagewidth=21
> >  >  >
> >  >  >  will override the system settings and set your pagesize to (1<<21)
> >  >  >  bytes.   I thought this would cause the madvise call to fail, but
> it
> >  >  >  doesn't.
> >  >  >
> >  >  >  Option needs renaming, but that is for later...
> >  >  >
> >  >  >  Please let me know if problems persist.
> >  >  >
> >  >  >  Take care,
> >  >  >  --
> >  >  >  Camm Maguire
> [email protected]
> >  >  >
> ==========================================================================
> >  >  >  "The earth is but one country, and mankind its citizens."  --
> Baha'u'llah
> >  >  >
> >  >  > _______________________________________________
> >  >  > Maxima-discuss mailing list
> >  >  > [email protected]
> >  >  > https://lists.sourceforge.net/lists/listinfo/maxima-discuss
> >  >  >
> >  >
> >  >  --
> >  >  Camm Maguire
> [email protected]
> >  >
> ==========================================================================
> >  >  "The earth is but one country, and mankind its citizens."  --
> Baha'u'llah
> >  >
> >  > _______________________________________________
> >  > Maxima-discuss mailing list
> >  > [email protected]
> >  > https://lists.sourceforge.net/lists/listinfo/maxima-discuss
> >  >
> >
> >  --
> >  Camm Maguire
> [email protected]
> >
> ==========================================================================
> >  "The earth is but one country, and mankind its citizens."  --
> Baha'u'llah
> >
> > _______________________________________________
> > Maxima-discuss mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/maxima-discuss
> >
>
> --
> Camm Maguire                                        [email protected]
> ==========================================================================
> "The earth is but one country, and mankind its citizens."  --  Baha'u'llah
>

_______________________________________________
Maxima-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/maxima-discuss