Seems like one really _needs_ to use CVS snapshot now... is something preventing release of version 5.6?

[Thomas Orgis]

Hi,

I just set up my new system with kindof current hardware (x86_64) and software (GNU/Linux, with xorg-server 1.5 and OpenMotif 2.3.1).
Regarding nedit I realized that I must use the CVS snapshot to prevent basic crashes like this:

ncl $NOT_YET_EXISTING_FILE

X Error of failed request: BadValue (integer parameter out of range for operation)
Major opcode of failed request: 91 (X_QueryColors)
Value in failed request: 0xffb3b3b3
Serial number of failed request: 5046
Current serial number in output stream: 5046

(I guess it's about the very dialog that asks about creating a new file, as opening existing files works.)

At least on the ever-evolving GNU/Linux platform, I have the feeling that nedit is becoming more and more endangered by obsolescense.
Apart from the Unicode issue (which is nagging), we really should get a release with the fixes to make nedit work with current versions of Xorg and Openmotif (dunno if there are still outstanding issues with 2.3.1 and nedit HEAD? Will see.).
Bitrot is unforgiving...


Alrighty then,

Thomas.


PS: Probably I should try to prepare a patch for those when I got some time:

macro.c: In function 'dialogMS':
macro.c:2927: warning: cast to pointer from integer of different size
macro.c: In function 'dialogBtnCB':
macro.c:2968: warning: cast from pointer to integer of different size
macro.c: In function 'stringDialogMS':
macro.c:3104: warning: cast to pointer from integer of different size
macro.c: In function 'stringDialogBtnCB':
macro.c:3158: warning: cast from pointer to integer of different size
macro.c: In function 'listDialogMS':
macro.c:3683: warning: cast to pointer from integer of different size

That's only a few examples... I guess we should have those fixed before it becomes common for nedit to indeed have a virtual address space above 2G (editing/viewing some 10-year apache log or so;-).

[Bert Wesarg]

Post by Thomas Orgis
macro.c:2927: warning: cast to pointer from integer of different size
macro.c:2968: warning: cast from pointer to integer of different size
macro.c:3104: warning: cast to pointer from integer of different size
macro.c:3158: warning: cast from pointer to integer of different size
macro.c:3683: warning: cast to pointer from integer of different size

I have done this before, and I have done it again now and checked all
these warnings, and its always the case that we assign a small integer
to a pointer and later cast the pointer back to an integer. Therefore
false-positives. And these integers are 'small', or will you have
sometime in the far future more that 2^32 menu entries or programming
languages? As far as I can see, the greatest number ever cast to a
pointer and back is (MAXPATHLEN + 1) (thats the size for the shell
name buffer in preferences.c).

The warnings in Xlt/SlideC.c, are only a problem, if on some
architectures NULL is something >=2^32, the same for
source/regularExp.c.

We have found some 64bit problems in the past and I use nedit for 5
years on a x86-64. So you can count on me that I'm behind this.

Regards,
Bert

PS: I've attached a patch to silent all warnings, execpt these for
Xlt/SlideC.c and source/regularExp.c.