Re: [PATCH] HAVE_MKSTEMP tries to kill not existent tmp-file in?$HOME and $TMPDIR (1.9.2)
Rado S <[email protected]> Fri, 25 May 2007 17:22:28 +0200
| Newsgroups | gmane.network.tin.devel |
|---|---|
| Message-ID | <[email protected]> |
=- Urs JanÃen wrote on Fri 25.May'07 at 16:27:01 +0200 -=
> > > sverrno = errno;
> > > if (fd == -1 && sverrno)
> >
> > The functionality hasn't changed,
>
> it now also checks the returnvalue of mkstemp, not only errno
Oh, right ... but there is also a final check for fd==-1 after all
the #ifdef, where it says "Couldn't create tmpfile...", have you
removed that? It would be redundant.
> > why have you put the extra sverrno var there?
>
> cause errno might be clobbered by the next system call and we need
> it's original val in the error message.
I understand the principle, but in this case there is no syscall
before it's being printed.
> {...} and for add_headers which is used to _log_ the message-id
> and the date header is the posting statistic.
Argh, I've misunderstood the code there then, sorry.
> > Anything else maybe? Isn't a MSG-ID a good thing to have?
>
> date and message-id are usualy inserted be the newsserver (the
> first server _must_ add them if the article dosn't contain them).
Fine ... but ... when tin reports "Article posted:" it shows a
msg-ID generated with the domain name of the server: does tin create
it that way or is this reported back from the server?
I was wondering why it didn't carry my host's domain.
--
© Rado S. -- You must provide YOUR effort for your goal!
EVERY effort counts: at least to show your attitude.
You're responsible for ALL you do: you get what you give.