HP-UX : problems with multiple calls to process-send-region

HP-UX : problems with multiple calls to process-send-region

[andrew.maguire]

I have a subprocess to which I am sending data using process-send-region.
I have found that with HP-UX 11 the subprocess is apparently not receiving
(a middle) part of the data.
As an experiment, I advised process-send-region to call sit-for with a small
amount of time before each call. If the time is set to > 0.01 seconds
all the data is successfully passed. With delays < 0.01 an increasing amount
of data
is being lost.

I am using Emacs 21.2 but I have checked that the problem existed in a
patched
version of 20.3 too also.

Any suggestions/solutions?

TIA

Andrew

Re: HP-UX : problems with multiple calls to process-send-region

[Kai Großjohann]

andrew.maguire@ps.ge.com writes:

> I have a subprocess to which I am sending data using process-send-region.
> I have found that with HP-UX 11 the subprocess is apparently not receiving
> (a middle) part of the data.
> As an experiment, I advised process-send-region to call sit-for with a small
> amount of time before each call. If the time is set to > 0.01 seconds
> all the data is successfully passed. With delays < 0.01 an increasing amount
> of data
> is being lost.
>
> I am using Emacs 21.2 but I have checked that the problem existed in a
> patched
> version of 20.3 too also.
>
> Any suggestions/solutions?

Interesting.  A Tramp co-developer (producing more code than me) also
found this problem recently.  It exists in the CVS version of Emacs,
too.  It's not clear to me what can be done about it.  Tramp includes
a workaround (involving sleep-for I think), but of course it would be
better to fix the C code.

Do you know something about HP-UX development and could you have a
look at the Emacs code to see what's wrong?  I know little C and less
HP-UX, so it's difficult for me...
-- 
A turnip curses Elvis

RE: HP-UX : problems with multiple calls to process-send-region

[andrew.maguire]

I do little C too. But I will try to find some time to investigate this
further.

I also found that if I built Emacs on HPUX 11.00 then the cursor would
only flash when Emacs received mouse or keyboard input.

I found that if I built Emacs on an HPUX 10.20 box the cursor flashed as
it should do.

Andrew

> -----Original Message-----
> From: kai.grossjohann@uni-duisburg.de
> Interesting.  A Tramp co-developer (producing more code than me) also
> found this problem recently.  It exists in the CVS version of Emacs,
> too.  It's not clear to me what can be done about it.  Tramp includes
> a workaround (involving sleep-for I think), but of course it would be
> better to fix the C code.
> 
> Do you know something about HP-UX development and could you have a
> look at the Emacs code to see what's wrong?  I know little C and less
> HP-UX, so it's difficult for me...

> andrew.maguire@ps.ge.com writes:
> 
> > I have a subprocess to which I am sending data using 
> process-send-region.
> > I have found that with HP-UX 11 the subprocess is 
> apparently not receiving
> > (a middle) part of the data.
> > As an experiment, I advised process-send-region to call 
> sit-for with a small
> > amount of time before each call. If the time is set to > 
> 0.01 seconds
> > all the data is successfully passed. With delays < 0.01 an 
> increasing amount
> > of data
> > is being lost.
> >
> > I am using Emacs 21.2 but I have checked that the problem 
> existed in a
> > patched
> > version of 20.3 too also.
> >
> > Any suggestions/solutions?
>