Enrico Forestieri wrote:
On Fri, Dec 14, 2007 at 07:23:32PM +0100, Abdelrazak Younes wrote:
Enrico Forestieri wrote:
On Fri, Dec 14, 2007 at 06:51:34PM +0100, Abdelrazak Younes wrote:
Enrico Forestieri wrote:
On Fri, Dec 14, 2007 at 04:17:49PM +0100, Abdelrazak Younes wrote:
Jean-Marc Lasgouttes wrote:
I am trying to see why the UserGuide makes errors when typesetting. I
already fixed a crash, but it is still not working as I would want.
Anyway, is there a reason why I see on the console the messages below?
It makes debug messages basically useless.
Hum, probably my fault. The original copy() used to overwrite by
default. There is an "overwrite" option to copyTo(), I guess we should
pass it true when moving files from the Cache. I'll do that.
There's another problem with the new FileName::renameTo() method.
The old rename() method was basically equivalent to move:
OK, then I'll add a moveTo() method and use that.
No, IMO the old behavior should simply be restored.
Too late, sorry. The old behavior _is_ restored. This is only about
naming things along what they do. If system level rename() does not
overwrite files, neither should we. And if we need something similar to
the system level 'move' then we should name this method moveTo(). I
don't really understand your resistance. If you are afraid of new bugs,
don't worry, I'll fix them in due time. We are not talking here about
methods that are used all other the place...
I don't fear bugs that show up. I fear bugs that don't show up, meaning
bugs that is difficult to spot because they only manifest themselves
in subtle ways not directly related to the changed semantics.
Well, I think subtle bugs are often a result of unclear semantics. So by
getting rid of obscure semantics, we often get rid of bugs. A function
name should be crystal clear as to what it does and do it right; we
should not modify the called method just to turn around problems in the
callers.
Abdel.
