although creating objects with variable case works fine on all plattforms, what doesn't work is the help patches - at least on windows. for example, if you look for help from [Line~], you'll get [line~] instead. would it be possible to correct this to make it more cohese? or is the help search plattform dependent, and thus impossible to fix?

On 2010-04-25 15:04, Mathieu Bouchard wrote:
On Sun, 25 Apr 2010, Frank Barknecht wrote:

Just a quick note: as operating systems treat upper- and lowercase in
filenames differently (Linux is picky and will not load
renderDepthSelector.pd for an object called [renderdepthselector],
while Windows and OS-X will do), I try to avoid uppercase in
abstraction names.

On Windows and OSX, [renderdepthselector] won't create either, though
the file will be open. GetProcAddress and dlsym are case-sensitive. Then
after that, class_new is case-sensitive too (because selectors are
case-sensitive).

all true, but GetProcAddress will never be called for a file called
"renderDepthSelector.pd".


anyhow, i (personally) don't think that CamelCase objectnames are bad.
just make sure that you really use them throughout, both on the
filesystem level and on the patching level.
(or even better: switch to a case sensitive filesystem in the first
place :-))


Or you go with operating system features: On *nix systems you could
use a symbolic link like "ln -s renderDepthSelector.pd rds.pd".

Only on filesystems that support it. I say that because FAT is becoming
*more* common on those operating systems; so when you transmit the file
using flash-memory formatted the default way, you have to tar it or zip it.


i didn't know that zip would handle symbolic links. (never tried though)
thanks for the idea.

fgmasdr
IOhannes



--
Friedenstr. 58
10249 Berlin (Deutschland)
Tel +49 30 42020091 | Mob +49 162 6843570
Studio +49 30 69509190
[email protected] | skype: jmmmpjmmmp

_______________________________________________
[email protected] mailing list
UNSUBSCRIBE and account-management -> 
http://lists.puredata.info/listinfo/pd-list

Reply via email to