Racket 5.3 bugs

[Ras Bodik]

Hi all,

 

If you downloaded Racket recently, you probably got the 5.3 version.  We have already seen two bugs in this new version.  In particular, our sim-syn code does not work correctly. Please download the 5.2 version, which seems to work fine.

 

--Ras

[Emina Torlak]

Hi all,

The second bug that Ras mentions is fairly serious if you are communicating with an external solver process via its standard input stream.  The problematic behavior happens when there are two threads running concurrently:  one writing the formula to the solver's input stream, and another writing output to a completely unrelated port (say, debugging info to Racket's own standard output).  In this scenario, Racket 5.3 will occasionally get the wires crossed and write fragments of the unrelated output to the solver's input stream.  This happens when the formula is relatively large (in my case, ~1400 lines of encoding). I cannot reproduce it with small problems, and I have not seen this problem with version 5.2.1.

Emina

--