Re: Standard libc (Was: Re: Sorry)
On Wed, 19 Aug 1998, Alan Cox wrote:
> > was just wondering from where you got your list of things that should be
> > in libc. I haven't had the time to look through it entirely, but I do see
>
> Its generated from the Libc documentation.
I see now. The stuff in dlfcn.h isn't very well documented (I don't think
I could find an info page if my life depended on it).
> > that you are missing the functions from dlfcn.h (the library for dynamic
> > loading/unloading of libraries). I admit that I am personally biased
> > towards these functions (dlopen, dlclose, dlsym, and dlerror) being
> > included in the standard library set because of a project I am working on,
>
> Fine. Send me a diff for the HTML document for stuff I've missed
I will do my best. It may take a while, though, as I will try to though
all of glibc2 by hand. Yech.
> > standard, as they are already part of glibc2, and are also part of UNIX98.
> > IMHO, it is important for a Linux standard to meet and/or exceed the
> > minimums of UNIX(tm) branding, even if Linux itself never gains official
> > branding. My $0.02.
>
> The goal is to keep the standard small, but I agree witht he dlopen/dlclose
> etc). Its an oversight not a decision.
I agree that it is important to keep the standard small. I have sitting in
front of me here the 27 pages of system calls that are listed as either
mandatory or part of a 'Feature Group' for UNIX 98 compliance (listed one
per line, no descriptions). I think that I should find most of the
functions listed on UNIX 98 on your list already, though my list doesn't
specify #defined constants as yours does, so some things may slip by.
At least the list of commands necessary for UNIX 98 is only 5 pages long.
:)
+-----------------------------+--------------------------------+
| Jakob 'sparky' Kaivo | jake@nodomainname.net |
| NoDomainName Networks | http://www.nodomainname.net |
+-----------------------------+--------------------------------+
Reply to: