jimm@amc.com (Jim McElroy) (05/18/91)
I am using a version of cfront 2.1 on a 386 MS-DOS system. It uses a "DOS extender" that arranges for the program to execute in extended memory, in the '286 16mB mode. In most respects it works fine, but it fails on switch statements that have more than about 15 cases. The failure reports that stack overflow has occurred -- during the execution of cfront. Can someone with access to the source tell me if "cfront" uses large amounts of stack space when compiling switch statements? Is it reasonable for me to "lean" on my compiler vendor to fix this? By the way, the switch statement is used in low-level code (lexical analyzer) where virtual function invocations would not be appropriate. Thanks in advance for any info. ---- Meta-Analytics - Software for Embedded Systems Jim McElroy Voice: (206) 747-4349 FAX: (206) 747-4353