This is the mail archive of the newlib@sources.redhat.com mailing list for the newlib 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: [PATCH]: Implement getprogname and setprogname


cgf@redhat.com wrote:
On Mon, Nov 17, 2003 at 08:08:05PM -0500, Nicholas Wourms wrote:

Hi all,

Corrina's latest check-in for stdlib.h prodded me into finishing the documentation for my implementation of these two functions for newlib (as opposed to just Cygwin). So, I'm submitting a patch with source, inline texinfo documentation, and the bsd manpage. I am also working on quite a few other bsd/SuSv3/c99 functions which seem to be located in most libc's "gen" dir. So I hope you don't mind that I created that dir. More contributions to follow.

Also, if you approve it, please commit the bzipped patch, not the inlined text one, since my MUA is munging tabs and some whitespaces.


I suspected that you might chime in on this, but before you jump to conclusions, please hear me out.


Then you should use another MUA.

I don't see what the fuss is all about, after all I did include an inlined txt version for commentary. As for switching, I'm afraid that's not possible, since I'm not changing freemail providers and netscape.net requires that I use Netscape (not just mozilla) to get imap access to my mail. I suppose I could just send it from a local exim daemon, so I guess I'll look into doing that in the future.



Code is based off of {Free,Net}BSD sources.


This is entirely BSD specific. IMO, it doesn't belong in newlib.

Indeed they are BSD specific, which was entirely my intent and I see no crime in it. As for them not belonging in newlib, you have a right to your opinion, however you also had plenty of opportunity to chime in on this thread:


http://sources.redhat.com/ml/newlib/2003/msg00509.html

But seeing as you didn't and since the discussion generally indicates otherwise, I'll assume that they do belong and they are wanted. Next up is my version of the *err/*warn functions from BSD, which benefit from having these two functions available.

In case anyone thinks I just copied the sources from Corinna's version in Cygwin, I didn't. I had already let Corinna know that I had working, newlib versions of these functions and the err/warn functions back when she committed the err/warn functions to Cygwin. The reason I hadn't already submitted them was because I hadn't finished my testing and fully documenting them at that time. I asked her what I should do, and she told me just to submit my work to newlib and if it was acceptable she would transition Cygwin to use it instead. I see this as a win/win situation, as it consolidates pure C implementations from Cygwin to newlib and extends newlib's core functionality for all platforms. These aren't just off-the-wall functions, they are actually quite frequently used.

Finally, let me add that I purposely didn't make the functions compiled by default for embedded targets in the case they are trying to achieve a minimal libc. If you or anyone else think I should set the ELIX level higher, then I'll be happy to comply. However, as far as I'm concerned, I see no reason newlib can't be both as robust as glibc or as lean as people want it to be. The great part about {free,opensource}-software is choice, and that's all I'm trying to offer here. My code is seperated and compartmentalized, so worries about it being extra bloat or unmanageable are frankly bogus. In any event, the prior interest for these and other BSD functions was indicated in that thread, which is what I intend to provide.

Cheers,
Nicholas


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