Hi Hans-Peter,
it looks like the problem has disappeared in the latest GNU APL version:
*
)LOAD MT631_DDD.apl
DUMPED 2025-04-20 18:25:18 (GMT+2)
)SAVE
2025-04-20 18:41:34 (GMT+2) MT631_DDD
*
Could you please check if the problem persists after a clean build, i.e.
*
make clean all
sudo make install*
Background:
There was a short period after your previous error report where I had to
revert to an
older cloning scheme (as controlled by Macro *NEW_CLONE* in file
*Value.hh*) to fix the
problem that you reported. After that period I returned to the new
cloning scheme.
The first stack trace below can only occur if the old cloning scheme was
in effect on *)SAVE*.
This makes me believe that your interpreter was built during that period.
Best Regards,
Jürgen
On 4/13/25 22:34, Hans-Peter Sorge via Bugs and suggestions for GNU APL
wrote:
Hi,
The Stack Trace is persistent if I try to )SAVE.
)DUMP, )LOAD, )SAVE does not reproduce the trace.
All Variables were )ERASEd
* )FNS*
ANALYSE_READINGS_UPDATE ANALYSE_REDINGS_UPDATE BYTES DTSs
EDIT H2D H2I H2N H2T ISK LO PROCESS RA
RAl RAr REC REC_CHR REC_CHR_BYTES RR SAVE SV TEST ∆∆
/⍝ The two functions //*ANALYSE_READINGS_UPDATE* and
*ANALYSE_REDINGS_UPDATE*
⍝ have identical content.
⍝ //*ANALYSE_REDINGS_UPDATE* had A typo! After multiple updates I
opened it
⍝ changed the function name to //*ANALYSE_READINGS_UPDATE*//and closed
the function
⍝ then )SAVE the WS and getting the Stack Trace.
⍝ Next I )ERASEd /ANALYSE_READINGS_UPDATE - the newly created function./
⍝ Stack Trace again/
*)ERASE ANALYSE_READINGS_UPDATE*
*)DUMP*
2025-04-13 17:04:42 (GMT+2)
*)SAVE*
*** Sub-Value 0xa8d866f0 has two parents.
Child: vid=-1, _val=0x2000000200, _par=11217
Parent 2: vid=699, _val=0xa9bbc1b0_par=-1
Call stack:
----------------------------------------
-- Stack trace at Archive.cc:1129
----------------------------------------
0x7fc10352a30b __libc_start_main
0x7fc10352a248
0x403e80 main
0x5dc75b Workspace::immediate_execution(bool)
0x46062b Command::process_line()
0x460807 Command::process_line(UCS_string&, std::ostream*)
0x46164b Command::do_APL_command(std::ostream&, UCS_string&)
0x4727f7 Command::cmd_SAVE(std::ostream&, UCS_string_vector const&)
0x5de93d Workspace::save_WS(std::ostream&, LibRef, UCS_string
const&, bool)
0x40f370 XML_Saving_Archive::XML_Saving_Archive(std::ostream&,
std::ostream&, char const*)
0x4138f7 XML_Saving_Archive::save()
========================================
========================================
Running )CHECK...
OK - no stale functions
OK - no stale values
OK - no stale indices
==============================================================================
Assertion failed: vvp
in Function: cmd_CHECK
in file: Command.cc:797
C/C++ call stack:
----------------------------------------
-- Stack trace at Assert.cc:75
----------------------------------------
0x7fc10352a30b __libc_start_main
0x7fc10352a248
0x403e80 main
0x5dc75b Workspace::immediate_execution(bool)
0x46062b Command::process_line()
0x460807 Command::process_line(UCS_string&, std::ostream*)
0x46164b Command::do_APL_command(std::ostream&, UCS_string&)
0x4727f7 Command::cmd_SAVE(std::ostream&, UCS_string_vector const&)
0x5de93d Workspace::save_WS(std::ostream&, LibRef, UCS_string
const&, bool)
0x40f370 XML_Saving_Archive::XML_Saving_Archive(std::ostream&,
std::ostream&, char const*)
0x413955 XML_Saving_Archive::save()
0x46411c Command::cmd_CHECK(std::ostream&, UCS_string const&)
0x423f1d do_Assert(char const*, char const*, char const*, int)
========================================
======== =============================== =
SI stack:
==============================================================================
*** immediate_execution() caught other exception ***
*)FNS*
ANALYSE_REDINGS_UPDATE BYTES DTSs EDIT H2D H2I H2N
H2T ISK LO PROCESS RA RAl RAr REC REC_CHR
REC_CHR_BYTES RR SAVE SV TEST ∆∆
/⍝ After erasing the function with just the missing character *"A"* /
⍝ The )SAVE is OK!
*)ERASE ANALYSE_REDINGS_UPDATE
* *)save *
2025-04-13 17:08:21 (GMT+2) MT631_EEEE
I hope the stack Trace has information to relate the cause to the
function renaming.
Other than that I have no idea what else might have caused the stack
trace.
The attached )DUMPed WS has both ANALYSE_READINGS_UPDATE and
ANALYSE_REDINGS_UPDATE
included. As I mentioned, )LOAD, )SAVE will not reproduce the bug.
May be some of the content
pops up in the trace.
The attached )SAVEed WS ends in a nirvana.
Best Regards
Hans-Peter
