Re: Cygwin patches

[Lennart Borgman]

On Mon, Nov 23, 2009 at 2:37 AM, Ken Brown <address@hidden> wrote:
> On 11/22/2009 6:30 PM, Lennart Borgman wrote:
>>
>> On Mon, Nov 23, 2009 at 12:23 AM, Ken Brown <address@hidden> wrote:
>>>
>>> On 11/22/2009 6:03 PM, Chong Yidong wrote:
>>>>
>>>> Ken Brown <address@hidden> writes:
>>>>
>>>>> Thanks for the suggestion.  My revised patch is attached.  If it looks
>>>>> OK, could someone please check it in?
>>>>
>>>> Your patch looks incorrect.  It's not your job to add the "file://"
>>>> part; that's done in the rest of the function.  Does this corrected
>>>> patch do the right thing on cygwin?
>>>
>>> Yes, it does do the right thing on cygwin.  As I said earlier in the
>>> thread,
>>> the rest of the function *doesn't* add "file://"; it adds "file:" without
>>> the slashes.  But if I add "file://", the rest of the function leaves
>>> this
>>> alone.  I guess browse-url-filename-alist would have to be changed to
>>> make
>>> it correctly add "file://" in this situation, but I haven't tried to dig
>>> into it to figure out how it works.  It seemed easier (and harmless) to
>>> just
>>> add it myself.
>>
>>
>> We were recently discussing if it should be file: or file:/// for w32.
>> I really do not know which is best but I suspect file:/// have a
>> slightly better chance to work. I just saw an instance
>> (css-validator.jar with Rhino) where it worked better.
>
> So are you saying that "file:" (with no slashes) is added by design?  It
> doesn't work for me.  (After the call to cygpath, I have a path of the form
> C:/blah/blah/blah.html; I need to add "file://", since otherwise only
> "file:" gets added and I don't have a valid URL.)


Yes, it is by design. Maybe it is because know one knows how many / to
add? In my case, plain w32, it should be three
(file:///c:/some/where.html).