On 22/10/2012 16:25, Alan Coopersmith wrote: > On 10/22/12 08:07 AM, Jon TURNEY wrote: >> From: Ryan Pavlik <rpavlik-KgB/ygh5px6vc3sceru...@public.gmane.org> >> >> Provide a dummy readlink, which just returns the supplied path name, on WIN32 >> >> Signed-off-by: Ryan Pavlik <rpavlik-KgB/ygh5px6vc3sceru...@public.gmane.org> >> Reviewed-by: Jon TURNEY <jon.turney-grjqepx9rppajuda+fb...@public.gmane.org> >> --- >> src/fontfile/catalogue.c | 9 +++++++++ >> 1 files changed, 9 insertions(+), 0 deletions(-) >> >> diff --git a/src/fontfile/catalogue.c b/src/fontfile/catalogue.c >> index 8029b28..7ec06f2 100644 >> --- a/src/fontfile/catalogue.c >> +++ b/src/fontfile/catalogue.c >> @@ -37,6 +37,15 @@ static const char CataloguePrefix[] = "catalogue:"; >> >> static int CatalogueFreeFPE (FontPathElementPtr fpe); >> >> +#ifdef WIN32 >> +static int readlink(const char *path, char*buf, size_t bufsiz) >> +{ >> + /* Silly no-op implementation */ >> + strncpy(buf, path, bufsiz); > > Don't you want a buf[bufsiz - 1] = '\0' after that, just in case > the caller passed something either not null terminated or too long > to fit the null byte in the buf?
I'm not sure. I can't find anything that says readlink() is supposed to behave like that, and indeed the man page I have on linux for readlink explicitly says it does not null terminate buf. I think this patch is superseded, anyhow. _______________________________________________ xorg-devel@lists.x.org: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel