Re: Stable CVS Version 1.11.13 Released!

[Paul Edwards]

"Derek Robert Price" <derek@ximbiot.com> wrote in message 
news:mailman.2434.1076870948.928.bug-cvs@gnu.org...
> >I wish people didn't hold on so dear to tabs (in source
> >code), they're a real pain.
>
> I'd actually be up for changing them in the CVS source since the mix of
> spaces and tabs makes it impossible to reset tab stops usefully anyhow,
> but I'd prefer going to all tabs with a default tabstop of 4, which
> would look the same.  :)

And then we'll still have whitespace problems when posting
patches.  Why not configure your editor to convert spaces
to tabs and back again, if you like editting source with tabs?
So long as the source code ends up without tabs, all will be
fine, there will be no problems like we're experiencing now.

> As for your cutting and pasting, one of your programs probably converted
> the tabs into spaces and at first glance it looks like this patch has
> the same problem.  I know that using netscape as an emailer and a diff
> viewer, cutting and pasting from the browser window to the email window
> preserves the tabs.  I'm not sure what other combinations work.

I don't have Netscape.  I just know how to use the basics of
Outlook Express.

Actually, I've just had a look, there appears to be an "insert"
option.  Let's see how this goes:

Index: update.c
===================================================================
RCS file: /cvs/ccvs/src/update.c,v
retrieving revision 1.226
diff -c -r1.226 update.c
*** update.c 2 Feb 2004 21:10:28 -0000 1.226
--- update.c 16 Feb 2004 02:43:15 -0000
***************
*** 2129,2137 ****
      {
   char *mrev;

-  if (rev2 != NULL)
-      free (rev2);
-
   /* If the first revision doesn't exist either, then there is
             no change between the two revisions, so we don't do
             anything.  */
--- 2129,2134 ----
***************
*** 2180,2185 ****
--- 2177,2185 ----
   {
       if (rev1 != NULL)
    free (rev1);
+      if (rev2 != NULL)
+          free (rev2);
+
       return;
   }

***************
*** 2206,2221 ****

       if (rev1 != NULL)
    free (rev1);

       return;
   }

!  /* If only one join tag was specified, and the user file has
!            been changed since the greatest common ancestor (rev1),
             then there is a conflict we can not resolve.  See above for
             the rationale.  */
!  if (join_rev2 == NULL
!      && strcmp (rev1, vers->vn_user) != 0)
   {
       if (jdate2 != NULL)
    error (0, 0,
--- 2206,2228 ----

       if (rev1 != NULL)
    free (rev1);
+      if (rev2 != NULL)
+          free (rev2);

       return;
   }

!  /* If the file has been changed, genuinely changed,
!            in comparison to the greatest common ancestor (rev1),
             then there is a conflict we can not resolve.  See above for
             the rationale.  */
!         if ((strcmp (rev1, vers->vn_user) != 0)
!             && RCS_cmp_file(vers->srcfile,
!                             rev1,
!                             (char **)NULL,
!                             rev2,
!                             vers->options,
!                             finfo->file))
   {
       if (jdate2 != NULL)
    error (0, 0,
***************
*** 2231,2242 ****
--- 2238,2253 ----

       if (rev1 != NULL)
    free (rev1);
+      if (rev2 != NULL)
+          free (rev2);

       return;
   }

   if (rev1 != NULL)
       free (rev1);
+  if (rev2 != NULL)
+      free (rev2);

   /* The user file exists and has not been modified.  Mark it
             for removal.  FIXME: If we are doing a checkout, this has

If this doesn't work, I'll post the patch as a hex dump.  I have
a C program to do a hexdump and to do the reverse.  That
should do the trick.

Please reconsider using tabs, they are a real pain.

BFN.  Paul.