Date: Mon, 16 Mar 1998 14:27:43 +0200 (IST) From: Eli Zaretskii To: Martin Stromberg cc: djgpp-workers AT delorie DOT com Subject: Re: Where to get the latest sources for djtar In-Reply-To: <199803161212.NAA22497@propus.lu.erisoft.se> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Precedence: bulk On Mon, 16 Mar 1998, Martin Stromberg wrote: > > I think testing for a full disk is too expensive, and also doesn't > > cover all the cases (e.g., a write-protected floppy that isn't full). > > Hardly too expensive, it's only done when an error occurs. But in some cases in point (a write-protected disk, for example) errors happen all the time, and we don't want djtar to bail out right away, so you will see quite a few of the errors and statfs it every time. > Isn't some better diagnostics worth the time? This diagnostics will be printed anyway if the disk fills while writing a file (as opposed to creating a directory). Maybe djtar should also test if the disk is full before it actually begins to work, or before creating each directory? > 24 slashes: this is what I empirically determined on my two partitions. I > couldn't find any #define for this. Does anybody know if this is defined > somewhere? I mean a #define in a header file as well as some DOS-ish FAT > definition. I think the limitation is not 24 slashes, it's the length of the path name. Anything longer than 64 characters should make DOS barf. But the problem is that this limitation is not consistently implemented (some DOS functions will succeed even if you exceed the limits), and Windows 95 (and probably 32-bit File Access in Windows 3.11) don't have these limitations. So I'm not sure if we should depend on this. > + /* See if its to deep directory nesting, in which case > + we should return ENAMETOOLONG - AMS */ > + if(_truename(dirname, tn)) `_truename' will give you hell and hight water if the drive is mounted via a network. I think it's better to use `_fixpath', if you need that.