V gets confused when dealing with Optional Sum Types and spits invalid C code

[RpxdYTX]

Describe the bug

When using ?SumType, the compiler sometimes gets confused and generates invalid C code. I found two scenarios where this happens.

Reproduction Steps

1. Casting ?T to U when T = U | Other:

type Sum = int | string
fn sum() ?Sum { return Sum(5) }

a := sum()
println(a as int) // V compiler should error, since we're casting from ?Sum to int

2. Assigning T to ?T and checking if ?T is U (when T = U | Other:

mut a := sum() // a: ?Sum
// ...
a = sum() or { ... } // (sum() or { ... }) converted to ?Sum
if a !is string { ... } // V compiler should error, since we're checking if ?Sum is string

Expected Behavior

The V compiler should error on both situations, preventing the error from propagating into the generated C code, which is harder to analyse

Current Behavior

In my current codebase, it gave these errors:

================== C compilation error (from cc): ==============                                                              cc:        |                                       ~~                                                                            ^~~~~~~~~                                                                                                                  cc: /data/data/com.termux/files/usr/tmp/v_10246/hex.01JGMQ477ZJQZGF038M50E0775.tmp.c:11955:51: error: no member named '_main__Ident' in 'struct _option_main__Token'
cc:  11955 |                 string name = (*(main__Ident*)__as_cast((ident)._main__Ident,(ident)._typ, 111)).ident;
cc:        |                                                         ~~~~~~~ ^
cc: /data/data/com.termux/files/usr/tmp/v_10246/hex.01JGMQ477ZJQZGF038M50E0775.tmp.c:11955:72: error: no member named '_typ' in 'struct _option_main__Token'
cc:  11955 |                 string name = (*(main__Ident*)__as_cast((ident)._main__Ident,(ident)._typ, 111)).ident;
cc:        |                                                                              ~~~~~~~ ^
cc: 3 warnings and 2 errors generated.
... (the original output was 17 lines long, and was truncated to 8 lines)
================================================================
(You can pass `-cg`, or `-show-c-output` as well, to print all the C error messages).
builder error:
==================
C error found. It should never happen, when compiling pure V code.

The code contained both cases in it.

Possible Solution

No response

Additional Information/Context

No response

V version

V 0.4.9 7b9b3dd

Environment details (OS name and version, etc.)

V full version: V 0.4.9 b487986.7b9b3dd OS: termux, 3.18.140-SIMPLE-KERNEL_V1.1, #1 SMP PREEMPT Sun Apr 18 03:22:59 -03 2021
Processor: 8 cpus, 64bit, little endian getwd: /storage/emulated/0/Documents/hex vexe: /data/data/com.termux/files/home/.v/v vexe mtime: 2024-12-31 18:09:35 vroot: OK, value: /data/data/com.termux/files/home/.v VMODULES: OK, value: /data/data/com.termux/files/home/.vmodules
VTMP: OK, value: /data/data/com.termux/files/usr/tmp/v_10246 Git version: git version 2.47.1 Git vroot status: weekly.2024.53 (15 commit(s) behind V master)
.git/config present: true
CC version: clang version 19.1.6 emcc version: N/A
thirdparty/tcc status: thirdparty-unknown-unknown de82a13

Note

You can use the 👍 reaction to increase the issue's priority for developers.

Please note that only the 👍 reaction to the issue itself counts as a vote.
Other reactions and those to comments will not be taken into account.

[JalonSolov]

You V is at least 15 commits behind the current. Do v up and try your examples again.

When I tried your first example here, it worked fine, with no errors from V or C. I saved it as x.v:

$ v run x.v
5
$

[RpxdYTX]

Sorry, i forgot the question mark before Sum in the sum function. Just upped v and it still yields C errors, instead of signalizing that converting from ?T to U is invalid, though the second example seems to have been solved

[jorgeluismireles]

The C error is produced when an unwrapped option is tried to be converted with as int:

type Sum = int | string
fn sum() ?Sum { return Sum(5) }

a := sum()?
println(a as int) // 5
b := sum()
println(b)        // Option(Sum(5))
println(b? as int) // 5
println(b as int) // C error

In the last line b is an unwrapped option converted to as int that V couldn't check.

Running code...
Can't run code. The server returned an error:
/tmp/v_60000/../../../../../../box/code.v:9: error: field not found: _int
builder error: 
==================
C error found. It should never happen, when compiling pure V code.
This is a V compiler bug, please report it using `v bug file.v`,
or goto https://github.com/vlang/v/issues/new/choose .
You can also use #help on Discord: https://discord.gg/vlang .
Exited with error status 1
Please try again.

See https://play.vlang.io/p/5a99d599af

[RpxdYTX]

So just

a := sum()
println(a as int)

doesn't trigger the error?

[RpxdYTX]

Just tested, it does

[jorgeluismireles]

Yes, simply these two lines:

type Sum = int | string
println(?Sum(5) as int)

[spytheman]

type Sum = int | string
println(?Sum(5) as int)

Please file another issue for this.

[RpxdYTX]

Isn't it the same thing?

[JalonSolov]

Doesn't appear to be. Just tried with latest update, and got

$ v -g run x.v
/tmp/v_1000/../../../../../../home/jalon/x.v:2: error: field not found: _int
builder error: 
==================
C error found. It should never happen, when compiling pure V code.
...