This is the mail archive of the mailing list for the cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Serious performance problems (malloc related?)

On Thu, Jun 02, 2005 at 03:55:50PM -0400, Igor Pechtchanski wrote:
>On Thu, 2 Jun 2005, Christopher Faylor wrote:
>> On Thu, Jun 02, 2005 at 12:39:17PM -0700, Shankar Unni wrote:
>> >Christopher Faylor wrote:
>> >>I am leery of doing things this way since that means that the only
>> >>people capable of writing code for cygwin are the people who understand
>> >>Nt* calls.  That is a subset of the already small number of people who
>> >>understand the UNIX and Windows APIs well enough to work on Cygwin.
>> >
>> >You mean, like 2?  (Err, 3.  I mean, 4, or maybe 5..)
>> >
>> >I'm not sure this is a huge problem, you know..
>> Just take a look at the number of people who have contributed to
>> cygwin-patches in the last year or so.  I wouldn't want to scare away
>> the people who contribute trivial patches because they can't find any
>> documentation on "NtCreateFile".
>> I suppose *we* could produce documentation on the Nt* functions but
>> that's also a support burden.
>If the Cygwin team had to write 9x wrappers for the Nt* functions, some
>documentation would have to be part of the wrappers, wouldn't it?

The theory is that Nt knowledgeable pepole could maintain a backend Nt*
layer and everyone else could write to the standard Win32 API vs.
exposing the Nt layer to everyone and requiring that people figure out
the unfamiliar Nt* arguments if/when they want to make changes.

But, Corinna and I haven't finished discussing how this would work yet
so there are no guarantees how or if this will happen.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]