On Tue, May 23, 2006 at 01:09:36AM +0200, Michael Kerrisk wrote:
> > reference ftw.3; see #367846
>
> Done. For 2.33, And also the converse SEE ALSO in ftw.3
> referring to fts.3.
>
> > Mention the return values in a separate section (especially fts_open)
>
> This page derives from BSD. Other than trivial changes, or
> major flaws, I'm not inclined to gratuitously make changes.
>
> > s/NULL-terminated/null-terminated/
>
> No! It is not a string!
| A single buffer is used for all of the paths of all of the files in the
| file hierarchy. Therefore, the fts_path and fts_accpath fields are guar-
| anteed to be NULL-terminated only for the file most recently returned by
| fts_read(). To use these fields to reference any files represented by
| other FTSENT structures will require that the path buffer be modified
| using the information contained in that FTSENT structure's fts_pathlen
| field. Any such modifications should be undone before further calls to
| fts_read() are attempted. The fts_name field is always NULL-terminated.
Both of these are strings.
> > Figure out what is up with FTS_{LOGICAL,PHYSICAL}; it seems that
> > flags combinations which are noted as being invalid (neither set)
> > and implied as invalid (both set) are not rejected. Bug#367691
> > and #333160 are mildy relevant (to me anyway, since I'm updating
> > the pages)
> >
> > Michael, I'm waiting for your comments on the patch suggested in
> > #367846 before changing this page.
>
> This note is the first mention I have seen of that patch.
>
> I have no interest in patch that changes multiple things
> at one time, without any explanation. If you want to
> get any of those changes in, do them as logically separate
> pieces (separate bug reports), and give them to me one at
> a time, along with as well reasoned argument. I suggest
> that you space the reports so that I can respond to one
> before you submit the next.
document FTW_ACTIONRETVAL;
include .SH "RETURN VALUE";
reorganized and rearranged the page
I'm not sure I can break the third one up, and it won't be well
represented in a diff. I think the page reads much more clearly now,
anyway.
Justin
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
