Ethereal-dev: Re: [Ethereal-dev] BUG: printing in windows needs access-rights to root director
On Friday 29 July 2005 10:05, Ulf Lamping wrote:
> Ph. Marek wrote:
> >Trying to print in windows results an error, as I'm working as a user and
> > the root-directory is not writable for me.
>
> As you are writing to the developer mailing list I assume you are a
> developer yourself.
diff -u print_dlg.c.orig print_dlg.c
--- print_dlg.c.orig 2005-07-29 10:26:17.363031600 +0200
+++ print_dlg.c 2005-07-29 10:30:14.314310680 +0200
@@ -855,7 +855,11 @@
win_printer = TRUE;
/*XXX should use temp file stuff in util routines */
g_free(args->file);
- args->file = g_strdup(tmpnam(NULL));
+ args->file = gstrdup_printf("%s\\ethereal_print_%s.tmp",
+ getenv("TMP") ? getenv("TMP") :
+ getenv("TEMP") ? getenv("TEMP") :
+ "c:\\temp",
+ tmpnam(NULL));
args->to_file = TRUE;
#else
g_free(args->cmd);
> In this case you should know that's not a good idea
> to only provide an interpretation of an error message, but provide the
> exact error message itself.
>
> You should also explain the exact steps to reproduce your problem. There
> is more than one printing function, so we have to guess which one you
> tried.
Using the "official" 0.10.12 ethereal build, capture some packets.
Choose File->Print, option don't matter, as long as you don't print into a
file :-)
On clicking OK I get a messagebox
Ethereal
You don't have permission to create or write to the file "\s1bg.".
This is with TEMP and TMP set to a writeable path (below profiles).
And in all cases in windows ./gtk/print_dlg.c:862 (the tmpnam()-call you
mentioned) is used, so it didn't seem necessary to me to specify any
particular options I used.
> >I couldn't find an option for setting the path; please use %TEMP% or
> > %TMP%.
> The function tmpnam() is used to get the output path/filename for
> printing on windows. The function description is a bit hard to
> understand, but I think it will use %TMP%, so there seems something
> wrong with your environment.
I don't think so - on both accounts (it doesn't, and there isn't ;-).
> >(BTW: could I have a 0.10.12 with that changed, please ? :-)
> No. We won't ship two different versions with the same 0.10.12 version
> number. That's simply a very bad idea.
How about a 0.10.12b, then ? :-)
> Regards, ULFL
Regards,
Phil
