>> "m. allan noah" <kitno455 at gmail.com> wrote: >> >>> True enough, So, instead of using mktemp at all, why not do something >>> like append '.part' to the image name? I think that is a little more >>> clear to the end user if they find some hanging around after a >>> scanimage crash or something. >> >> Absolutely, especially if this all happens in the working directory. > > It has to be in the same directory like the target files or rename() could > fail as I understand it. But of course it's not really the current working > directory because output filespec can include a path, relative or > absolute. > My intention for using mktemp() was to make sure that one scanimage > process does in any case rename only it's own file, even if a concurrent > scanimage process (maybe handling a different device) uses the same output > file name. > > Anyway I'll remove mktemp() tomorrow.
Updated patch is attached, hope that's less confusing now. Simon -------------- next part -------------- A non-text attachment was scrubbed... Name: sane-backends-scanimage-batch-atomic-output-files.patch Type: text/x-patch Size: 2197 bytes Desc: not available Url : http://lists.alioth.debian.org/pipermail/sane-devel/attachments/20090115/6fb01bee/attachment.bin
