Variables in a loop are overwritten

[hogsy]

The following code in a binary template is not possible.

...

while( !FEof() )
{
    set_comment_at("We're here...");

    // chunk header, basically tag and size
    struct GroupHeader chunkHeader;
    local uint32_t endOffset = FTell() + chunkHeader.size;

    char chunkData[chunkHeader.size];

    // seek to the end marker, just in case we did anything earlier
    FSeek(endOffset);

    // validate it has an end marker
    uint32_t chunkEnd;
    if ( chunkEnd != GROUP_CHUNK_END )
    {
        set_comment_at( "Hit invalid end!" );
        break;
    }
}

The variables chunkHeader, chunkData and chunkEnd will appear at the end of the file instead. The initial comment set at the start of each iterate will be set correctly however.

Expectation was that there would be an entry for those mentioned above added per loop to the file.

I can work around this by using locals and setting comments manually but obviously not ideal.

[solemnwarning]

I've come up with a couple of ways to do this, but not sure which would be best. Assuming a template like the following:

int num;

for(int i = 0; i < num; ++i)
{
    int foo;
}

1. We could mark each variable inside a loop with the offset where it was defined inside the comment, like so:

+------------------------+
| num                    |
+------------------------+
| 0000:0000  02 00 00 00 |
+------------------------+
+------------------------+
| foo @ 0000:0004        |
+------------------------+
| 0000:0004  00 00 00 00 |
+------------------------+
+------------------------+
| foo @ 0000:0008        |
+------------------------+
| 0000:0008  00 00 00 00 |
+------------------------+

2. We could define a comment encapsulating any loop which has variables within it, like this:

+----------------------------+
| num                        |
+----------------------------+
| 0000:0000  02 00 00 00     |
+----------------------------+
+----------------------------+
| Loop at template.bt:3      |
+----------------------------|
| +------------------------+ |
| | foo (0)                | |
| +------------------------+ |
| | 0000:0004  00 00 00 00 | |
| +------------------------+ |
| +------------------------+ |
| | foo (1)                | |
| +------------------------+ |
| | 0000:0008  00 00 00 00 | |
| +------------------------+ |
+----------------------------+

I'm leaning towards 1, since the latter could get weird with nested loops, but open to suggestions.

[hogsy]

I'd probably lean towards the first too.