JOE - Joe's own editor

Hi,

> It seems that there's still a problem with deleting the last character
> when wrapping the line. Even in the last 2.9.7pre2 SRPM. I attach the
> patch that works for me. There's a probability that it's not proper, but
> it really works and I haven't seen any side effects.
	I have tested this patch against 2.9.7pre2.tgz and it removes last
character when wordwrapping. After adding getc(q) (or it was getc(q)) into
uformat.c the problem with deleting last character ends. But problem with
word-wrapping remains :(
					marx

On Mon, 22 Oct 2001, Marek Grac wrote:

Hi Marc,

> > It seems that there's still a problem with deleting the last character
> > when wrapping the line. Even in the last 2.9.7pre2 SRPM. I attach the
> > patch that works for me. There's a probability that it's not proper, but
> > it really works and I haven't seen any side effects.
> 	I have tested this patch against 2.9.7pre2.tgz and it removes last
> character when wordwrapping. After adding getc(q) (or it was getc(q)) into
> uformat.c the problem with deleting last character ends.

No. I had to add both - getc(q) and my patch to not delete the last
character. My patch works against SRPM - 2.9.7pre2 with applied patches I
have found in SRPM. Addind only getc(q) call in uformat.c dows _not_ work
for me.

> But problem with word-wrapping remains :(

What problem? Cutting the line in the wrong place?

regards
pkot
-- 
mailto:pkot@... :: mailto:pkot@...
http://kt.linuxnews.pl/ :: Kernel Traffic po polsku
http://tfuj.pl/cv.html :: http://tfuj.pl/pgp.asc

Hi there,

On Mon, Oct 22, 2001 at 12:38:46 +0200, Pawel Kot wrote:

> > But problem with word-wrapping remains :(
> What problem? Cutting the line in the wrong place?

Yes, exactly that problem.

I haven't tested 2.9.7pre2 yet, but I'm sure it still exists.

Incorrect wrapping is a real pain in the a^Wbehind, and makes joe
almost unusable! Can it be that hard to track it down? Apparently. My
efforts with ddd failed, because I cannot understand what the
uncommented source code does and don't know what the variables are
for...

Greetings,
Moritz

On Mon, Oct 22, 2001 at 12:11:37PM +0200, Marek Grac wrote:
> Hi,
> 
> > It seems that there's still a problem with deleting the last character
> > when wrapping the line. Even in the last 2.9.7pre2 SRPM. I attach the
> > patch that works for me. There's a probability that it's not proper, but
> > it really works and I haven't seen any side effects.
> 	I have tested this patch against 2.9.7pre2.tgz and it removes last
> character when wordwrapping. After adding getc(q) (or it was getc(q)) into
> uformat.c the problem with deleting last character ends. But problem with
> word-wrapping remains :(
> 					marx

	Hi!

  Does anyone know when this problems started? If we can find offending
change, we can revert it. Same applies to wordwrapping problem.

	Cheers,
		Vita

Hi guys,

On Mon, Oct 22, 2001 at 12:47:59 +0200, Vitezslav Samel wrote:

>   Does anyone know when this problems started? If we can find offending
> change, we can revert it. Same applies to wordwrapping problem.

Yes. It happened from 2.9.6pre0 to 2.9.6pre1 (or perhaps pre2?). I could
recompile each vanilla version to check, though.

I never knew where to look for the problem. In the meantime, I have
run all code versions through GNU-indent to make the source comparable,
but to no avail.

Best greetings,
Moritz

Hi!

  I made the diffing between the versions 2.9.6-pre1 and 2.9.6-pre2 and hope
I found the proper solution to the wordwrap problem (it also solves the
deleting of the last char on the previous line). Please, test enclosed path
(against joe-2.9.7-pre2), and I will release 2.9.7-pre3 soon.

	Enjoy it,

		Vita


diff -urN 01-abs_link/utils.c 02-wordwrap/utils.c
--- 01-abs_link/utils.c	Mon Jul  2 23:23:02 2001
+++ 02-wordwrap/utils.c	Mon Oct 22 16:41:50 2001
@@ -13,6 +13,18 @@
 #include <ctype.h>
 
 /*
+ * Define function isblank(c) for non-GNU systems
+ * !!! code which uses isblank() assumes tested char is evaluated only once
+ */
+#ifndef __USE_GNU
+int isblank(int c)
+{
+        return((c == 32) || (c == 9));
+}
+#endif
+
+
+/*
  * return minimum/maximum of two numbers
  */
 unsigned int uns_min(unsigned int a, unsigned int b)
diff -urN 01-abs_link/utils.h 02-wordwrap/utils.h
--- 01-abs_link/utils.h	Mon Jul  2 23:23:02 2001
+++ 02-wordwrap/utils.h	Mon Oct 22 16:40:34 2001
@@ -34,7 +34,7 @@
  * Define function isblank(c) for non-GNU systems
  */
 #ifndef __USE_GNU
-#define isblank(c) (((c)==32) || ((c)==9))
+int isblank(int c);
 #endif
 
 /*

On Mon, 22 Oct 2001, Moritz Barsnick wrote:

Hi,

> efforts with ddd failed, because I cannot understand what the
> uncommented source code does and don't know what the variables are
> for...

Completly agreed. The code is ununderstandable. And that is why my code
was "seeking in the dark"...

pkot
-- 
mailto:pkot@... :: mailto:pkot@...
http://kt.linuxnews.pl/ :: Kernel Traffic po polsku
http://tfuj.pl/cv.html :: http://tfuj.pl/pgp.asc