Fix crash from an incomplete `do-while` statement

[Daniel-Cortez]

What this PR does / why we need it:

Fixes the compiler crash from an incomplete do-while statement without the while (...); part (see #678). Function test() (file sc1.c) was saving and restoring the previous value of sc_intest through PUSHSTK_I()/POPSTK_I(), and this was preventing the compiler from stopping the compilation process when the end of file was reached before closing the do-while statement. Making the function save/restore the value of sc_intest using the actual stack (by saving the value into a local variable) fixes the issue.

Which issue(s) this PR fixes:

Fixes #678

What kind of pull this is:

• [x] A Bug Fix
• [ ] A New Feature
• [ ] Some repository meta (documentation, etc)
• [ ] Other

Additional Documentation:

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity.

[Y-Less]

Merged locally.

[Y-Less]

I get an access violation here after merging this PR:

/* pc_readsrc()
 * Reads a single line from the source file (or up to a maximum number of
 * characters if the line in the input file is too long).
 */
char *pc_readsrc(void *handle,unsigned char *target,int maxchars)
{
  return fgets((char*)target,maxchars,(FILE*)handle);
}

Exception thrown at 0x77A0FF05 (ntdll.dll) in pawncc.exe: 0xC0000005: Access violation writing location 0x000000B8.

I can't see why, looking at the change, but it was definitely introduced by 3e0c5ba9c0ae9ff3c86269010237c356d6bbd1fc

[Y-Less]

When compiling the YSI test script:

https://github.com/pawn-lang/YSI/blob/master/gamemodes/YSI_TEST.pwn

[Daniel-Cortez]

My bad, it seems I forgot to replace the use of POPSTK_I here:

https://github.com/Daniel-Cortez/pawn-3.10/blob/f053b83a48f81287c2871a5b00445fb102b73f86/source/compiler/sc1.c#L6020

This caused corruption of the value stack, would only happen when the control expression in if, while, do..while, for, assert or state results in a constant value, which is pretty rare as normally the compiler warns about this. In your case it was assert(false) in YSI_Players\y_groups\y_groups_entry.

[Daniel-Cortez]

OK, now it should be fixed, both here and in #681. Not sure if I should add a separate test for this, as the case seems to be very specific.