<br><br>Le mardi 15 juillet 2014, Ben Hoyt <<a>benhoyt@gmail.com</a>> a écrit :<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>
Victor had one other question:<br>
<br>
> What happens to name and full_name with followlinks=True?<br>
> Do they contain the name in the directory (name of the symlink)<br>
> or name of the linked file?<br>
<br>
I would say they should contain the name and full path of the entry --<br>
the symlink, NOT the linked file. They kind of have to, right,<br>
otherwise they'd have to be method calls that potentially call the<br>
system.<br>
</blockquote><div><br></div><div>Sorry, I don't remember who but someone proposed to add the follow_symlinks parameter in scandir() directly. If the parameter is added to methods, there is no such issue.</div><div><br>
</div><div>I like the compromise of adding an optional follow_symlinks to is_xxx() and stat() method. No need for .lstat().</div><div><br></div><div>Again: remove any garantee about the cache in the definitions of methods, instead copy the doc from os.path and os. Add a global remark saying that most methods don't need any syscall in general, except for symlinks (with follow_symlinks=True).</div>
<div><br></div><div>Victor</div>