[GRASSLIST:4346] Re: .gislock woes & success

[Bill Sneed]

Glynn Clements wrote:


> AFAIK, it has also been reported with some versions of gcc-2.96. 
> However, the problem invariably manifests itself as an "internal
> compiler error", triggered by some dubious code (a function which is
> declared as returning "double" but doesn't actually include any
> "return" statements).
> 
> Certain code-generation switches cause the problem to disappear, as
> does changing the function's return type to "void".
> 
> 


Thanks for info...

Unfortunately I have no simple way to test 5.0.0pre5 with gcc2.96 (don't 
want to uninstall gcc3.1 !) but I never experienced this problem with 
previous versions of either GRASS or gcc on this hardware (which is 
unchanged across multiple versions of GRASS)

If you think the "error" message from trying gmake5 -i etc, etc, would 
be of help to someone, I'll pass it along....it's beyond my meager 
abilities to dig through the code....

...bill sneed, prospect, maine...