This is the mail archive of the cygwin mailing list for the Cygwin project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Cygwin process handling

We had discussions about this back in May and a solution was put in place and it has helped. However, I am still seeing intermittent hanging when GNU Emacs (emacs-w32) is interacting with back-end GNU Linux systems.

I run bash shell sessions in GNU Emacs buffers and SQL sessions on remote hosts thru Emacs buffers as well.  I see progress hanging on save operations, using the "with-editor" MELPA add-in (opens a Emacs buffer tied to a remote file from within a Emacs bash session), and remote logins created to start a sqlplus session.

I send a SIGUSR2 to the Emacs session to get it's attention, and the backtrace generally looks like ("xyzmachine" is an obvious replacement for the actual server name):
  accept-process-output(#<process *tramp/scp mmauger@xyzmachine*> 1 nil t)

  byte-code("\306^H\206^F^@\307\310\311#\211^Y\nB^Z\312\216\313^K^H\f^M\205^Y^@\314$+\207" [timeout -with-timeout-timer- with-timeout-timers p timeout-msecs proc run-with-timer 1 nil #[nil "\300\301\211\"\207" [throw timeout] 3] ((cancel-timer -with-timeout-timer-)) accept-process-output t] 6)
  tramp-accept-process-output(#<process *tramp/scp mmauger@xyzmachine*> 1)

  tramp-wait-for-regexp(#<process *tramp/scp mmauger@xyzmachine*> nil "\\(^\\|^@\\)[^#$\n]*///c9cdf3c04442cbe5185fcf9c69da1204#\\$^M?$")
  tramp-wait-for-output(#<process *tramp/scp mmauger@xyzmachine*>)

Timeouts don't seem to get it's attention.  Sometimes C-g C-g C-g works, but not always. I realize I need to be more specific but it is happening only occasionally and usually when I'm under work deadline pressure. This is happening on my work PC (my home PC runs Fedora and I've seen no such behavior, but I don't do as much remote processing in that environment) so recompiling and debugging options are limited.

My intent was only to make others aware that I am still seeing some unexpected behavior in a fairly normal GNU Emacs work flow, and hope that someone else can pipe in with "Me, too!" and "Here's how to fix it!" :)

Let me know if there is anything else I can do to help you understand the issue further.

MICHAEL MAUGER // FSF Member // GNU Emacs sql-mode maintainer // GNU Linux, GNU Emacs, OwnCloud

Please no MS Word attachments (I don't pay MS; they don't own my computer)

Problem reports:
Unsubscribe info:

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]