Re: fatal: internal error Ver 5.2.0 readall()

[Andrew J. Schorr]

Hi,

FYI, this used to work properly.  When I checkout gawk as of commit
508e9368d2d461b2290af6787cad6dfbed357176, this error does not occur:

sh-5.1$ ./gawk -l extension/.libs/rwarray.so -f /tmp/write.awk
writeall returns 1

sh-5.1$ ./gawk -l extension/.libs/rwarray.so -f /tmp/read.awk
arr1: len=0
readall returns 1
arr1: len=0

According to git bisect, I introduced the problem in subsequent commit
07d04b91d1ac3e19fda5e029f1a244ea714c1d13. I need to dig in and figure out what
I broke.

Regards,
Andy

On Wed, Sep 14, 2022 at 07:37:17AM -0400, Andrew J. Schorr wrote:
> Ah, OK, that makes sense. I will get to this stuff, but it may take a couple
> of months.
> 
> Regards,
> Andy
> 
> On Wed, Sep 14, 2022 at 12:39:24AM -0600, arnold@skeeve.com wrote:
> > I did see it when I looked closer. I didn't copy paste the reader
> > program correctly the first time.
> > 
> > In any case, calling readall() as the first thing works correctly.
> > 
> > Andy, I will leave the namespace stuff for you to resolve. I think
> > you will have to look for names of the form foo::bar and split them
> > apart when installing the variable, but I don't remember for sure.
> > 
> > Thanks,
> > 
> > Arnold
> > 
> > "Andrew J. Schorr" <aschorr@telemetry-investments.com> wrote:
> > 
> > > I expected you to say that. But I'm just curious -- do you not see
> > > the same error I'm seeing? You said above that it does not reproduce,
> > > so I'm confused by that.
> > >
> > > I will dig into this at some point.
> > >
> > > Regards,
> > > Andy
> > >
> > > On Tue, Sep 13, 2022 at 01:10:27PM -0600, arnold@skeeve.com wrote:
> > > > So Andy, you now "own" rwarray.c; please put this on your TODO list.
> > > > 
> > > > Thanks,
> > > > 
> > > > Arnold
> > > > 
> > > > "Andrew J. Schorr" <aschorr@telemetry-investments.com> wrote:
> > > > 
> > > > > Hi,
> > > > >
> > > > > I'm seeing the same issue as OP on Linux CentOS 7 using the master 
> > > > > branch:
> > > > >
> > > > > bash-4.2$ ./gawk -l extension/.libs/rwarray.so -f /tmp/write.awk
> > > > > writeall returns 1
> > > > >
> > > > > bash-4.2$ ./gawk -l extension/.libs/rwarray.so -f /tmp/read.awk
> > > > > arr1: len=0
> > > > > readall returns 1
> > > > > gawk: /tmp/read.awk:5: fatal: internal error: file interpret.h, line 
> > > > > 253: unexpected parameter type Node_val
> > > > >
> > > > > This sadly seems like a bug to me. If the code instantiates arr1 
> > > > > before
> > > > > caling readall, then the saved arr1 in the state file sohuld be 
> > > > > ignored
> > > > > if all goes according to plan. I'm not sure I have time to 
> > > > > troubleshoot
> > > > > right now.
> > > > >
> > > > > Regards,
> > > > > Andy
> > > > >
> > > > > On Sun, Sep 11, 2022 at 01:12:02PM -0600, arnold@skeeve.com wrote:
> > > > > > Hi.
> > > > > > 
> > > > > > In the future, please send source files as attachments, and not with
> > > > > > output mixed in to the middle; it makes it easier for me to deal 
> > > > > > with
> > > > > > things.
> > > > > > 
> > > > > > It turns out that this does reproduce on Linux.
> > > > > > 
> > > > > > Changing the reader program to be:
> > > > > > 
> > > > > > @load "rwarray"
> > > > > > BEGIN {
> > > > > >     ret = readall("foo.binawk");
> > > > > >     print ret " = readall(\"foo.binawk\");"
> > > > > >     printf("arr1: len=%s\n",length(arr1))
> > > > > > }
> > > > > > 
> > > > > > causes things to work without trouble. In short, you should not
> > > > > > force arr1 to be an array using split, but let readall handle 
> > > > > > setting
> > > > > > up the types and the contents.
> > > > > > 
> > > > > > Thanks,
> > > > > > 
> > > > > > Arnold
> > > > > > 
> > > > > > J Naman <jnaman2@gmail.com> wrote:
> > > > > > 
> > > > > > > # Sorry, I do not have a "gawkbug script"
> > > > > > > in a nutshell: readall() gives me 'fatal: internal error: file 
> > > > > > > interpret.h,
> > > > > > > line 253: unexpected parameter type Node_val' as described below:
> > > > > > >
> > > > > > > # TEST writeall()
> > > > > > > # Gawk ver 5.2.0 (mingw) api3.2 GNU MP 6.1.2 GNU MPFR 4.0.2
> > > > > > > @load "rwarray"
> > > > > > > BEGIN{
> > > > > > > split("",arr1); for(i in SYMTAB) { if(!isarray(SYMTAB[i])){
> > > > > > > arr1[i]=SYMTAB[i];}}
> > > > > > > ret = writeall("foo.binawk");
> > > > > > > print ret " = writeall(\"foo.binawk\");"
> > > > > > > exit;
> > > > > > > }
> > > > > > > # stdout: 1 = writeall("foo.binawk");
> > > > > > > # TEST readall()
> > > > > > > # Gawk ver 5.2.0 (mingw) api3.2 GNU MP 6.1.2 GNU MPFR 4.0.2
> > > > > > > @load "rwarray"
> > > > > > > BEGIN{
> > > > > > > split("",arr1);
> > > > > > > printf("arr1: len=%s\n",length(arr1))
> > > > > > > ret = readall("foo.binawk");
> > > > > > > print ret " = readall(\"foo.binawk\");"
> > > > > > > #**> fatal: internal error: file interpret.h, line 253: unexpected
> > > > > > > parameter type Node_val
> > > > > > > printf("arr1: len=%s\n",length(arr1))
> > > > > > > exit;
> > > > > > > }
> > > > > > > #===============
> > > > > > > # stdout: arr1: len=0
> > > > > > > # stdout: 1 = readall("foo.binawk");
> > > > > > > # foo_readAll.awk:12: fatal: internal error: file interpret.h, 
> > > > > > > line 253:
> > > > > > > unexpected parameter type Node_val
> > > > > > >  Note: granted one can use reada() and writea(), but when I got 
> > > > > > > the error
> > > > > > > with readall(), I cut out all variables until only the one array 
> > > > > > > was left
> > > > > > > that caused the error. Thanks, John
> > > > >
> > > > > -- 
> > > > > Andrew Schorr                      e-mail: 
> > > > > aschorr@telemetry-investments.com
> > > > > Telemetry Investments, L.L.C.      phone:  917-305-1748
> > > > > 152 W 36th St, #402                fax:    212-425-5550
> > > > > New York, NY 10018-8765
> > > > 
> > >
> > > -- 
> > > Andrew Schorr                      e-mail: 
> > > aschorr@telemetry-investments.com
> > > Telemetry Investments, L.L.C.      phone:  917-305-1748
> > > 152 W 36th St, #402                fax:    212-425-5550
> > > New York, NY 10018-8765
> > 
> 
> -- 
> Andrew Schorr                      e-mail: aschorr@telemetry-investments.com
> Telemetry Investments, L.L.C.      phone:  917-305-1748
> 152 W 36th St, #402                fax:    212-425-5550
> New York, NY 10018-8765

-- 
Andrew Schorr                      e-mail: aschorr@telemetry-investments.com
Telemetry Investments, L.L.C.      phone:  917-305-1748
152 W 36th St, #402                fax:    212-425-5550
New York, NY 10018-8765