Why static syntax checkers often fail for Tcl/SDC/UPF scripts?

If you’ve written Tcl/SDC/UPF scripts for chip design workflows, you’ve probably experienced this: your script runs for hours, only to crash due to a basic runtime error. It’s one of the most frustrating things in scripting.

Static checkers like dcprocheck are supposed to catch these issues early, but in practice, they often miss real problems — or worse, flag incorrect ones.

Here are three real examples I’ve encountered:

Example 1 — Unreachable and unchecked source files:

source bad_subscript_filepath.tcl
source good_subscript_filepath.tcl

Two problems here:

1. The missing file (bad_subscript_filepath.tcl) does not trigger an error in the static check.
2. The valid subscript (good_subscript_filepath.tcl) is not checked at all, even though it may contain real logic or syntax errors.

This also applies to read_sdc and load_upf, which may internally pull in scripts. If these aren’t actually parsed and validated, you're flying blind.

Example 2 — Missed runtime error:

set i -3 

while { 3 / $i } { # should FAIL after 3 iterations 
    puts $i 
    incr i 
} 

This should clearly fail after a few iterations (division by zero), but the static checker reports no error. The script crashes only at runtime.

Example 3 — False error:

if true { 
    puts "PASS" 
} else { 
    puts this line should not be checked 
} 

The else-branch should be ignored, yet the checker wrongly reports:

Line 4: E Wrong number of arguments of "puts"

What tool will actually help you?

TclOK — a Tcl/SDC/UPF script checker with static+dynamic checking mechanism.

For example 1&2, it reports the correct errors:

Error: couldn't read file "bad_filepath.tcl": no such file or directory 
Line: 2 
File: pro.tcl 

Error: divide by zero 
Line: 5 
File: pro.tcl

And for example 3, it ignored the dead branch entirely — as expected in real execution.

Learn More:

https://dashthru.com/tclok