Re: patch and PR bin/39326

[Joerg Sonnenberger <joerg%britannica.bec.de@localhost>]

On Sat, Sep 20, 2008 at 10:20:31AM -0400, Greg Troxel wrote:
>  Does anyone know if the problem I found (silent truncation
> when /tmp is too small) exists in the new version?
> 
> I looked at the sources, and don't see that ofp is checked with ferror
> after fputs in apply_hunk.  So perhaps this just needs an

It checks at the end if any error happened and complains if it did.
I tried to test this with tmpfs and I think I never seen an actual error
case happen. Note for review: fclose of input is not checked as errors
here are not of interest or have shown up earlier. When switching to
Plan B, it doesn't care about intermediate errors either. All other
cases check for ferror() before fclose().

Joerg