[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Bison 1.50 experiences
From: |
Hans Aberg |
Subject: |
Re: Bison 1.50 experiences |
Date: |
Tue, 15 Oct 2002 23:49:33 +0200 |
At 11:42 -0700 2002/10/15, Paul Eggert wrote:
>> Well, that probably depends on the compiler. Has that happened recently? --
>
>Sure. It happens on Solaris 9, for example. You can cause it to
>happen to any filesystem by setting TMPDIR to a filesystem that is
>nearly full.
Well, I am satisfied with using fixed names for the temporary files, as I
suggested. One can unlink them, and if they remain by some error, or as
they do anyway when requesting tracing, they will be easily identifiable.
>Anyway, it looks like we can redo Bison to not need any temp files at
>all, so the issue will be moot.
How should that work? It is M4 that is the problem: It needs to read a
sequence of files of which the Bison computed outputs is in the second
and/or third ones:
m4 ... m4sugar.m4 definitions skeleton
as in m4_invoke().
Hans Aberg
Re: ver 1.50 crashes, Akim Demaille, 2002/10/14