Re: ATTN: daemon maintainers -- proposed csih changes [Was: Re: [ITP] tftp-hpa 5.0]

[Charles Wilson <cygwin at cwilson dot fastmail dot fm>]

On 11/15/2010 2:12 PM, Corinna Vinschen wrote:
> On Nov 15 12:32, Charles Wilson wrote:
>> Obviously, the easiest solution is to just document the current behavior:
>>   ... with the exception of csih_is_nt2003, which true for both 64bit
>> Windows XP, as well as NT Server 2003 and higher, since XP64 shares the
>> same kernel as (64bit) NT2003.  See email thread blah for more
>> information...
>>
>> and update the messages about privileged users in the documentation and
>> warning messages (which needs to be done, regardless).
> 
> +1
> 
>> Really, this is what I get for violating autoconf's recommendations when
>> "designing" csih: never base capability tests on OS version numbers;
>> test for and indicate the capabilities themselves...
> 
> That's why we introduced wincap in Cygwin at one point...

Hmm...there's an idea for a utility: one that provides shell access to
the wincap values.  I'll think about that.

For now tho, I've implemented the change above (that is, no behavioral
change -- just updating the wording in csih's error/warning messages).
However, some of the foo-config scripts repeat the same "mistake" and
talk about "NT2003 or above" -- those scripts should also be updated to
reference XP64 as well, but it isn't critical.

I've published the csih-0.9.2-1 as a test release; it has the following
changes:

o Remove all vestiges of pre-NT support.
o Update user messages and comments to accurately reflect handling of
  Windows XP 64bit. Note that this means 'csih_is_nt2003' is now
  /documented/ to return true for Windows XP 64, in additional to
  the expected "Windows NT Server 2003 and above". This is necessary
  because XP64 shares the same kernel as NT2003, and thus both display
  similar behavior regarding the privileges of the built-in SYSTEM
  account.
o getVolInfo uses cygwin_create_path instead of deprecated
  cygwin_conv_to_full_win32_path.

There was one other pending csih change in my inbox...but it turns out
that problem is with Pierre's cron-config and exim-config, and not csih
after all.  I'll explain more in the appropriate thread.

--
Chuck