See patch. Feel free to add stuff/work on anything on the list. :)
>From 557d0b9b5008e05bf8dfe1fddd6d8d940158950d Mon Sep 17 00:00:00 2001
From: Pascal Stumpf <pascal.stu...@cubes.de>
Date: Wed, 24 Jul 2013 14:54:53 +0200
Subject: [PATCH] Add TODO.
---
cde/TODO | 79 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 79 insertions(+)
create mode 100644 cde/TODO
diff --git a/cde/TODO b/cde/TODO
new file mode 100644
index 0000000..a7fee2c
--- /dev/null
+++ b/cde/TODO
@@ -0,0 +1,79 @@
+General:
+* get rid of all remaining warnings. Goal is to be able to use -Wall without
+ too much annoying garbage in the build logs; especially:
+ * Replace strstream usage in ttsnoop
+ * LOTS of "string constant to char* conversions"
+ * unused variables
+ * 64bit problems. The most obnoxious cases are where pointers are converted
+ to ints and then used as array subscript. Currently worked around by using
+ longs, but of course, that is far from optimal. We need something better
+ (some sort of standardised hashtables?).
+* Usage of unsafe functions like strcpy, strcat, sprintf, mktemp, tmpnam etc.
+ The OpenBSD linker warns about those. They should be replaced, and, if
+ necessary for your OS, an implementation of safe alternatives (strlcpy etc.)
+ should be provided.
+* Likewise, some scripts have been using temporary files in an unsafe way.
+ This can mostly be avoided by using mktemp(1) - someone should check *all*
+ scripts.
+* Many times, it it assumed that malloc(3) zeroes allocated memory. Also, many
+ malloc(3) calls are vulnerable to int overflows (i.e. malloc(n * sz)). Both
+ problems can be avoided by using calloc(3). As these problems are somewhat
+ hard to detect, a complete audit should be performed ...
+
+* PREFIX cleanliness/FHS: Historically, /usr/dt has been the standard location
+ of CDE. This value is hardcoded in many places still. Find them and replace
+ with CDE_INSTALLATION_TOP.
+* Pending further discussion: Do not require a separate prefix for the CDE
+ hierarchy at all, but instead use customisable variables for BINDIR, SHAREDIR
+ etc.
+
+OS support:
+* CDE currently works on modern *BSD and Linux. I think we ought to keep the
+ possibility to extend that support to most other Unices, even the ones that
+ were dropped by the upstream vendor (IRIX, OSF ...). However, keeping
+ support for ancient versions of these OSes is debatable. Dropping those
+ might really serve to disentangle #ifdef mazes.
+* Hunt down abuses of processor macros (i.e. __ppc__, __sparc__) to test for
+ OS features
+
+Man pages / Documentation:
+* Japanese documentation does not build.
+* Section numbering follows SVR4 conventions. However, *BSD and most Linux
+ distributions follow BSD conventions. This should be configurable at build
+ time, changing e.g. 1m -> 8, 4 -> 5 ...
+* Many pages generate "table width larger than line length" warnings in groff.
+ Don't know if it's possible to work around.
+* For the crazy: Write a doctomdoc program to generate pages in a much nicer
+ language.
+
+Icons/Wallpapers/Artwork:
+* They pretty much suck on any halfway modern monitor. Provide better
+ resolutions. (Or better yet, somehow hack SVG support into CDE, haha).
+
+Printing:
+* Remove requirement for Xprint. It is obsolete, and many distributions do
+ not provide a Motif package linked against libXp any more.
+* Write a BSD lpr(1) backend. Currently, there is only support for the various
+ SVR4 lp(1) implementations, and BSD lpr(1) is substantially different.
+* Test CUPS as a backend. This is likely a lot easier than BSD'd LPD, but may
+ still need a few tweaks.
+
+DtMail:
+* Write support for lockspool(1) mailbox locking on OpenBSD (halfway done).
+
+DtKsh:
+* Investigate build against newer ksh93. This is problematic due to license
+ reasons (EPL is incompatible to (L)GPL).
+
+Dtinfo:
+* Get it working :)
+
+Dtsession:
+* Requiring suid root just for the screensaver SUCKS. Flesh out the
screensaver
+ bits into a smaller dtscreensaver program, perhaps, or just use xlock(1) or
+ xscreensaver(1) with customised CDE themes.
+
+OS-specific problems:
+* OpenBSD:
+ dtlogin(1): Keyboard does not work at login screen.
+ dtsession(1): Apparently, keyboard layout is borked in the screensaver.
--
1.8.3.1
------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
cdesktopenv-devel mailing list
cdesktopenv-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/cdesktopenv-devel
