On Sat, Mar 01, 2003 at 11:08:31AM +0100, Andreas Aardal Hanssen wrote:
> I'm doing some changes in the structure of the IO class/implementation,
> but if you post the patch I'm sure we can work it in.
Ooh, I'm glad you told me. Here's a quick and very dirty patch that is my
workaround for the bug in the IO class. I had planned on doing something
better before submitting it but if you are working in IO...
I had to put a fix in in order to test my COPY changes, otherwise
Binc was failing on incoming emails (from the client) using foreign language
or the full 8859-* character sets. (Finally a useful purpose for spam: testing)
--- bincimap-1.0.24/src/io/io.h 2003-03-01 02:00:59.000000000 -0800
+++ bincimap-1.0.24-dist/src/io/io.h 2003-02-21 12:01:10.000000000 -0800
@@ -235,7 +235,6 @@
//--
inline int readChar(void)
{
- int nextChar;
char c;
if (inputBuffer.size() != 0) {
@@ -248,10 +247,9 @@
return c;
}
- if (mode == IO::MODE_PLAIN) {
- nextChar = fgetc(stdin);
- c = nextChar;
- } else if (mode == IO::MODE_SSL) {
+ if (mode == IO::MODE_PLAIN)
+ c = fgetc(stdin);
+ else if (mode == IO::MODE_SSL) {
do {
int retval = SSL_read(ssl, &c, 1);
@@ -271,8 +269,8 @@
} while (1);
}
- if (nextChar == EOF)
- throw InputException("readChar EOF");
+ if (c == EOF)
+ throw InputException("readChar");
else {
if (c == '\n')
charnr = 0;
> I'm currently working on moving all of the Maildir code out of the main
> program files and into the Maildir class. Some improvements to the
> procedures of iterating over the message sets. The goal is to get
> everything so modular that contributors can stick to writing only a new
> MessageDepot child to support new backends. :-)
Sounds good. I'm getting tired so I'm going to give my -copy code another
once over and after killing my locked up mozilla session, try moving more
mail around and I'll send along the patch. I'll also send separate patches
for the changes I made to make testing easier. Specificallly turning off
setuid and setgid in the config file.
Do you prefer the patches as files on the web server or do you prefer
them on the mailing list? The longer patches are what I'm worried about,
at the moment the copy patch is 200 some lines long, though that will
be tightened up a bit as I discard some more commented out code.
C=)
--
--------------------------------------------------------------------------
Better the hard truth than the comforting fantasy. -- Carl Sagan
--------------------------------------------------------------------------
Caskey <caskey*technocage.com> /// TechnoCage Inc.
--------------------------------------------------------------------------
A presumption on your part does not constitute an obligation on my part.
