[M3devel] ***SPAM*** Re: Probably missing too much of development sycle, but WTF?

[Mika Nystrom]

These things are constant for a given binary distribution, aren't
they?  Can they be left to some sort of bootstrapping step on the
target system, or at least (if cross-compiling) in an environment
where target C headers are available?

I mean printing out an M3 interface based on the C header's defined
values.

    Mika

Jay writes:
>
>> AF_UNIX
> 
> 
>I'll try to make another pass through and put back unused constants, I guess.
> 
> 
>At some point I erred toward removing unused stuff, because everything present was a correctness and portability issue.
> 
>Though I mainly only did such removal with Cygwin, on the basis of it never really existing as a port so the initial version could be minimal.
>I think, maybe I removed other?
>I also removed userthread support from PPC_LINUX and maybe NetBSD/i386.
>It was broken for all platforms and remains so for all Linux platforms.
> 
> 
>Later I developed more portable techniques so it becomes more acceptable to leave stuff in, but convert it to be correct and portable.
> 
> 
>Constants are easy.
> 
> 
>> stuct_sockaddr_un structure
> 
> 
>Structs though are problematic.
>You end up wanting to wrap all the functions that use them.
>Unless they are large structs and then you don't want the waste and you are either stuck, or you maybe wrap at the next level up.
> 
> 
>This is an area actually where we still clone the headers, still work to do to make it portable and safe.
> 
> 
>> make_dir
> 
> 
>No I had forgotten but I see it searching the web.
>You need the dots smushed out?
> 
> 
> - Jay
>
>
>
>----------------------------------------
>> From: dragisha at m3w.org
>> To: jay.krell at cornell.edu
>> Date: Wed, 20 May 2009 07:41:34 +0200
>> CC: m3devel at elegosoft.com
>> Subject: Re: [M3devel] ***SPAM*** Re: Probably missing too much of development sycle, but WTF?
>>
>> Then please be complete :)
>>
>> AF_UNIX is missing, and also stuct_sockaddr_un structure.
>>
>> Yes I know how to add it. Yes you're pretty radical these days, so
>> please do it your way so you don't have to edit my patch :)
>>
>> Also - you maybe remember make_dir problem I've reported few months (?)
>> ago... I've narrowed it to situation where my earlier patch sets
>> CM3_INSTALL_PREFIX. It started to happen some months after I patched it,
>> and it happens only when shipping m3core.
>>
>> On Tue, 2009-05-19 at 14:37 -0700, jay.krell at cornell.edu wrote:
>>> Yes a little unfortunate but very portable.
>>>
>>> - Jay (phone)
>>>
>>> On May 19, 2009, at 10:27 AM, Dragi¹a Duriæ wrote:
>>
>>>
>>>
>>>> VARIABLES?
>>>>
>>>> Now every1 who was using CASE with error results has to adapt one's
>>>> code?
>>>>
>>>> What is rationale here? Portability?
>>>>
>>>> TIA!
>>>>
>>>> INTERFACE Uerror;
>>>>
>>>> FROM Ctypes IMPORT int;
>>>>
>>>> (* CONST *)
>>>>  VAR EACCES: int;
>>>>  VAR EADDRINUSE: int;
>>>>  VAR EADDRNOTAVAIL: int;
>>>>  VAR EAGAIN: int;
>>>>
>>>>
>>>> --
>>>> Dragi¹a Duriæ 
>> --
>> Dragi¹a Duriæ 
>>