On Fri, 7 Jun 2019 11:45:59 +0200 Zoran Vasiljevic <z...@archiware.com> wrote:
> At some time, if needed, the writer can be expanded > to understand Tcl channels as well. OK. Back-off. That would never work that way. Because the channel needs to be tossed back and forth, at lest 2 times. Once for the first write and once for the last. The first case is simple. The last isn't without user telling the system that he will not write any more (which requires additional "-last" option...) But even with this above solved... All this is also questionalble w.r.t feedback. How do I know anything went into the file? At some point one also has to introduce some kind of limit as otherwise we can blow memory by very fast producers and slow consumers... As I see, this could only work for "fire-and-forget" cases were one is not interested in any result and where one is producing small amount of data to be written. This is a highly specialized case, for example, for feeding log files an similar. But not for doing any kind of decisions based on the file or content of it... Now, being that specialized, one can also argue if a more general-purpose solution is justifed. Sorry for raining on the party, but it is good to know what/when/why *in advance* before putting too much work (or expectations) in the matter. _______________________________________________ naviserver-devel mailing list naviserver-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/naviserver-devel
