Re: [Nmh-workers] m_getfld

[David Levine]

Paul V. wrote:

> i'm taking that as approval by people who will variously do the work of:
> 
> 1. getting all non-rfc822-reading callers of m_getfld() to call
> something else;

That shouldn't be a blocker, see next point.

> 2. removing any m_getfld code that was just for those callers, even if
> it means removing arguments from all other calls to m_getfld;

I don't believe that there is any such code.

> 3. teaching m_getfld to never try to go behind the veil of stdio's api;
> 4. teaching m_getfld to stop optimizing for the vax pcc.

Would it make sense to toss it all and use mmap()?
Conceptually, it doesn't seem like a difficult problem to
parse header fields and maybe a body and just hand back a
pair of pointers and the state indication.

David