cmd/compile: SIGSEGV on NaCl when accessing implicit-zero elements in a large slice

[johnrichardrinehart]

The playground seems to trigger SIGSEGV runtime exceptions when accessing undefined elements in a large slice. I could not reproduce this locally (I get compile-time too-much-required-memory errors for large slices and normal runtime behavior for smaller slices).

What version of Go are you using (go version)?

Go 1.12.9

Does this issue reproduce with the latest release?

Not on a local machine, only on the playground.

What operating system and processor architecture are you using (go env)?

Native Client, NaCl.

What did you do?

I ran the following 3 programs.

1. https://goplay.space/#Rbv05s_S5ax
2. https://goplay.space/#Maqi_gYFakb
3. https://goplay.space/#NIfrZa_IIxb

What did you expect to see?

1. Not a SIGSEGV
2. I got what I expected.
3. I got what I expected.

What did you see instead?

1. A SIGSEGV runtime exception.
2. What I expected.
3. What I expected.

[agnivade]

I do get a seg fault when running locally, although not with the full stack trace.

GOOS=nacl GOARCH=amd64p32 go run foo.go 
signal: segmentation fault

[cuonglm]

I do get a seg fault when running locally, although not with the full stack trace.

GOOS=nacl GOARCH=amd64p32 go run foo.go 
signal: segmentation fault

You have to setup this https://github.com/golang/go/tree/master/misc/nacl for running nacl

[agnivade]

Woops, thanks @cuonglm. I never ran a basic program to check whether it works or not.

[bcmills]

Note that the NaCl port is on its way out anyway: see #30439 and #25224.

[johnrichardrinehart]

Oh, these are old issues. Sorry, @bcmills: I didn't do my due diligence. I'll go ahead and close this. No point in issuing a fix.