comp.lang.ada
 help / color / mirror / Atom feed
From: Mark Johnson <Mark_H_Johnson@Raytheon.com>
Subject: Re: gnat compiler -  help wanted
Date: Thu, 24 Jan 2002 08:38:20 -0600
Date: 2002-01-24T08:38:20-06:00	[thread overview]
Message-ID: <3C501C5C.F7862A15@Raytheon.com> (raw)
In-Reply-To: 9eb165bd.0201240338.14ea705a@posting.google.com

Tony wrote:
> 
> I am currently having problems making a server process which deals
> with the compilation of ada code from text delivered over a socket.
> 
I don't suppose you could...
 - create a temporary directory
 - feed the text from the socket to gnatchop (directly or indirectly),
output to that temporary directory
 - run gnatmake in that temporary directory
 - send the results back to the originator
Or is that considered "cheating"?

> [snip]
> My problem with the Gnat Source is following the dependencies.
> i.e. where does gnatmake come from... which packages (there are quite
> literally thousands of them).
> 
If you MUST figure out how gnatmake is built (which doesn't appear to be
a necessary precondition), I would suggest something as simple as...
  rm *.ali *.o
  gnatmake [switches] gnatmake 2>&1 | tee gnatmake.output
to rebuild gnatmake & capture the compiler output. Then edit the
gnatmake.output file to have the list of files that got compiled. I do
something similar when preparing an example to send to ACT for bug
reports.
  --Mark



  reply	other threads:[~2002-01-24 14:38 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-01-24 11:38 gnat compiler - help wanted Tony
2002-01-24 14:38 ` Mark Johnson [this message]
2002-01-24 15:05 ` Robert A Duff
2002-01-28 13:46   ` John English
2002-02-07 15:35   ` Tony Gair
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox