Summarize from a partial NNTP read

Summarize from a partial NNTP read

[Hallvard B Furuseth]

Sometimes the NNTP server is slow and I want to get at some articles
instead of waiting for more headers.

Sometimes the server hangs when it has sent 40-200 K headers.  (Well,
Gnus hangs anyway, I don't know what the server is doing.)

In both cases, I'd like to ^G and generate a *Summary* of the headers
that the server did manage to send.  Is that possible?

Today I ^G and try to fetch the group again, but this time with fewer
articles at a time.  So I must wait for the same headers to be fetched
all over again.

-- 
Hallvard

Re: Summarize from a partial NNTP read

[Lars Magne Ingebrigtsen]

Hallvard B Furuseth <h.b.furuseth@usit.uio.no> writes:

> Sometimes the NNTP server is slow and I want to get at some articles
> instead of waiting for more headers.
> 
> Sometimes the server hangs when it has sent 40-200 K headers.  (Well,
> Gnus hangs anyway, I don't know what the server is doing.)

Hm.  Well, it would be interesting to find out why that happens...

> In both cases, I'd like to ^G and generate a *Summary* of the headers
> that the server did manage to send.  Is that possible?

It actually wouldn't be that difficult to do -- one could just catch
the `C-g' (when fetching headers) and then just point Gnus at the
half-arrived headers.  Gnus assumes that any articles it can't get
headers for have expired, so one would have to do something about
that, which means that I think it's too much work anyway.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Summarize from a partial NNTP read

[Hallvard B Furuseth]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:
>Hallvard B Furuseth <h.b.furuseth@usit.uio.no> writes:
>> Sometimes the server hangs when it has sent 40-200 K headers.  (Well,
>> Gnus hangs anyway, I don't know what the server is doing.)
> 
> Hm.  Well, it would be interesting to find out why that happens...

Tell me what to debug and I'll try some day.

>> In both cases, I'd like to ^G and generate a *Summary* of the headers
>> that the server did manage to send.  Is that possible?
> 
> It actually wouldn't be that difficult to do -- one could just catch
> the `C-g' (when fetching headers) and then just point Gnus at the
> half-arrived headers.  Gnus assumes that any articles it can't get
> headers for have expired, so one would have to do something about
> that, which means that I think it's too much work anyway.

Maybe it would be simpler to not catch the C-g, and not clean up after
the partial fetch (until the next Gnus command, anyway).  Then the user
could call a command similar to `gnus-group-select-group' which would
use the previous NNTP data instead of rereading NNTP, and pretend it
only asked for the headers that were actually received.

-- 
Hallvard

Re: Summarize from a partial NNTP read

[Lars Magne Ingebrigtsen]

Hallvard B Furuseth <h.b.furuseth@usit.uio.no> writes:

> >> Sometimes the server hangs when it has sent 40-200 K headers.  (Well,
> >> Gnus hangs anyway, I don't know what the server is doing.)
> > 
> > Hm.  Well, it would be interesting to find out why that happens...
> 
> Tell me what to debug and I'll try some day.

`nntp-retrieve-headers' fetches the non-NOV headers, but I think it's
more likely that the bug is in `nntp-wait-for'.  

> Maybe it would be simpler to not catch the C-g, and not clean up after
> the partial fetch (until the next Gnus command, anyway).  Then the user
> could call a command similar to `gnus-group-select-group' which would
> use the previous NNTP data instead of rereading NNTP, and pretend it
> only asked for the headers that were actually received.

That would be a possibility, but it wouldn't be less work than
catching `C-g', really -- the same article number wrangling would have
to be done...

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen