Subject: Trivial actsync bug in INN 1.7.2 To: inn-bugs@isc.org Date sent: Sun, 11 Jan 1998 12:16:56 +0000 (GMT) From: USENET news manager A minor oddity that I just noticed with INN 1.7.2 (+insync 1.1d patches): Running actsync (on a test server, to sync it with routine additions and deletions on our live server) with options so it would prompt for each change, in the middle of a prompting sequence I got changegroup de.comm.mobil n [yn]? Must be running ctlinnd changegroup de.comm.mobiln exited with status: 1 The explanation for "Must be running" was that flushing the feeds had kicked off a c-nocem batch processing run, which then paused the server, so that's unsurprising. The trivial bug is in the line reporting the failed command: the group flag ("n") is tacked on immediately after the group name ("mobiln" instead of "mobil n"). Trivial, but worth tidying "in passing" when someone's working on actsync. A quick look at actsync.c suggests the problem is that in several blocks of error-handling code there are lines like (type ? "" : " "), (type ? type : ""), (who ? "" : " "), (who ? who : ""), exitval); as part of fprintf arguments, which are intended to output a space and then some text if the relevant text is non-null. Instead, they output a null string and then the text if it's set, and an unwanted space if the text is not set... I.e. the first line above should be (type ? " " : ""), (type ? type : ""), (" " and "" swapped), similarly for the "who" line, in each of the error-handling blocks where this consistently arises (a consequence of cut and paste, I assume!). John Line -- Cambridge University Computing Service - USENET news manager. Usually John Line newsmaster@ucs.cam.ac.uk (alias {newsmaster,news,usenet}@news.cam.ac.uk)