Re: [hylafax-users] Interesting problem with Hylafax+

[Lee Howard <faxguy@xxxxxxxxxxxxxxxx>]

Rob Poe wrote:

> Noticed a bug where faxqclean was fixed up to not remove cover pages that are being referenced.  I upgraded to the version of Hylafax+ that said it made those fixes (built from .SRPM).  Running version 4.3.0.12-1 downloaded from the SourceForge Hylafax+ site.  
>
> However, it seems to be still removing the cover page that HylaFax wants to send ..  The cover69.cover is being left because it's referenced by the job, but cover69;c1 is being removed

Thanks for bringing this up.  Attached is a patch to fix it.  You should 
be able to apply it to your SRPM from the spec file and rebuild.  This 
will be in HylaFAX+ 5.0.1.

Yes, there was a disagreement between faxq and faxqclean as to how the 
prepared images should be named.  faxq said "docq/cover123;f0" and 
faxqclean said "docq/cover123.cover;f0".  This patch brings faxqclean 
in-line with faxq.

Thanks,

Lee.
--- hylafax.orig/faxd/faxQCleanApp.c++	2006-11-13 22:37:28.000000000 -0800
+++ hylafax/faxd/faxQCleanApp.c++	2006-11-15 11:58:27.080482648 -0800
@@ -330,7 +330,8 @@
 	 * than the threshold (checked above).
 	 */
 	u_int l = file.nextR(file.length(), '.');
-	if (l != 0 && l < file.length()) {
+	u_int k = file.nextR(file.length(), '/');
+	if ((l != 0 && l < file.length()) || strncmp(&file[k], "cover", 5) == 0) {
 	    if (isdigit(file[l])) {
 		/*
 		 * Filename has a jobid suffix (or should);
@@ -404,7 +405,6 @@
 
 		free(base);
 	    }
-
 	}
 
 	if (nowork || Sys::unlink(file) >= 0) {