From: Alexandre F. <ale...@gm...> - 2016-07-24 13:40:46
|
Hi Donal, Can you please share some details ? First, which exotic OS exhibits the problem ? Though I admit I manage to keep a few parsecs clear from Windows these days, in the past I have heavily depended on pipes there, and never encountered more than userland bugs (e.g. in the Tcl channel machinery). But what you're saying here sounds like a kernel-side buffering that's completely alien to me. Note that the pipe buffering semantics can vary at the operating system levelsubstantially; it is not safe to assume that a write performed on the outputside of the pipe will appear instantly to the input side. This is afundamental difference and Tcl cannot conceal it. The overall stream semantics\fIare\fR compatible, so blocking reads and writes will not see most of thedifferences, but the details of what exactly gets written when are not. Thisis most likely to show up when using pipelines for testing; care should betaken to ensure that deadlocks do not occur and that potential short reads areallowed for. ---------- Forwarded message ---------- From: <aku...@sh...> Date: Sun, Jul 24, 2016 at 2:45 PM Subject: [Tcl-bugs] [Tcl Source Code] Commit by dkf - Make a few tests more resilient to differences in the semantics of pipes between operating systems. To: tcl...@li... |