Cyril Bouthors writes:
> 'GNU Make' could then complain with an explicit warning or
> error about internal cache and targets that confuse it.
If the message is explicit enough and includes pointer to the
documentation about internal cache, it will at least reduce the number
of questions ab
Patrick Craig writes:
>
> But you would only reread the directory when an error condition occured
> so Makfiles that worked before wouldn't be affected.
I agree with Paul that nobody wants to negates the purpose of
the cache but 'GNU Make' could do one more stat() just before exiting
"Paul D. Smith" wrote:
> %% Patrick Craig <[EMAIL PROTECTED]> writes:
>
> pc> I can understand using the directory caching feature to improve
> pc> performance, but it seems that this feature does cause users some
> pc> problems. Couldn't you modify make so that when such an error
> pc> o
%% Patrick Craig <[EMAIL PROTECTED]> writes:
pc> I can understand using the directory caching feature to improve
pc> performance, but it seems that this feature does cause users some
pc> problems. Couldn't you modify make so that when such an error
pc> occurs, instead of failing immediate
"Paul D. Smith" wrote:
> This doesn't have anything to do with suffixes; you could get the same
> behavior in other ways.
>
> The issue is GNU make's directory cache feature.
>
> When GNU make first needs to find a file in a directory it caches the
> contents internally. Then as it runs rules wh
%% Patrick Craig <[EMAIL PROTECTED]> writes:
pc> GNU Make version 3.79.1
pc> Built for i686-pc-cygwin
pc> My project has some autogenerated C files. I use a different
pc> suffix for these files (.agc) so that I can delete them in the
pc> clean rule (rm -f *.agc). Here is a simplified
