Variable setting regression in tcc17 ?

[Juppycmd]

I'm currently using TCC/LE, but as it's getting outdated I tried to test-drive the newest and fanciest version 17 (probably the very reason why there's no new LE anymore :-)).

Problem is: My batch files don't work no more :-( ... here's what happens:

TCC/LE 13:

[C:\Program Files\JPSoft\TCCLE13x64]set x=y
[C:\Program Files\JPSoft\TCCLE13x64]set %x%=z
[C:\Program Files\JPSoft\TCCLE13x64]echo %y%
z

TCC 17:

[C:\Program Files\JPSoft\TCMD17x64]set x=y
[C:\Program Files\JPSoft\TCMD17x64]set %x%=z
TCC: Not in environment "yz*"
[C:\Program Files\JPSoft\TCMD17x64]echo %y%
ECHO is OFF

Am I doing something wrong, or is nested var setting unsupported as of the last version?

 

[vefatica]

OK here.

v:\> type ifftest.btm
set x=y
set %x%=z
echo %y%

v:\> ifftest.btm
z

 

[Charles Dye]

I can reproduce it.

I think the issue almost certainly has to do with recent kludges to support pseudovariables. It'll work as expected if you either (A) disable pseudovariables, or (B) omit the (unnecessary) trailing percent sign after the variable name.

 

[rconn]

Charles is right - the %= is being interpreted as a pseudovar. If you don't use pseudovariables, it's much safer to turn them off in OPTION / Advanced.

 

[Juppycmd]

It'll work as expected if you either (A) disable pseudovariables, or (B) omit the (unnecessary) trailing percent sign after the variable name.

Thanks! The reason I use the trailing percent is that my editor (Notepad++) does nice color highlighting of the vars this way, and only this way. I didn't expect this to cause any trouble, or the Spanish Inquisition to arrive :-> ... but I'll be on my guard if similar errors show up.

 

[WavSlave]

NO ONE expects the Spanish Inquisition! (Sorry, I couldn't resist.)