Re: lynx-dev Batch makefile for MingW32 and Windows

[vtailor]

>>On Sun, 7 Nov 1999 address@hidden wrote:
>>
>>> Read 4 KB of data.  80.00 KB/sec.
>>>
>>> (It freezes after reading the first 4KB.)  If I push a key like `Q', it
>>> reads the next page:
>>>
>>> Read 8 KB of data.   0.32 KB/sec.
>>>
>>> etc., etc.
>>>
>>> If one of the developers knows where in lynx the code was changed so that
>>> not reading from the keyboard blocks something, let me know.
>>
>>Look at HTCheckForInterrupt in LYUtils.c.  I don't know, but maybe it
>>gets called in more situations now from HTFile.c, or from HTParseFile
>>or HTFileCopy (in HTFormat.c).
>>
>>Myabe you can work around some of the problem by diabling partial
>>display (-partial:off).  That's not a solution, of course.
>>
>>I would compare the 2.8.2 version of HTCheckForInterrupt with the
>>current one, esp. w.r.t. ifdef'ing.
>>
>>   Klaus
>
>Thanks for the suggestion.  Oh, BTW, there is also the problem that, once
>you start the new lynx, it doesn't call the Windows 98 dialer and it go
>into action if the dialup PPP connection is already working.

The problem was indeed in the function that you suggest:  It is an

#ifndef _WINDOWS
#ifdef USE_SLANG

that should read:

#ifdef _WINDOWS
#ifdef USE_SLANG

Once you correct the strange misspelling of ifndef above, you get a
working - and stable - lynx.exe for both slang and pdcurses.  I posted
the pdcurses version of November 3 source on alt.binaries.misc, just so
you can see this.

Of course, the mouse still does not work under pdcurses, but it does
work under slang.

Also, I posted elsewhere that the display corruption problem that occurs
when calling viewers can be traced to XSystem.c and its implementation
of system_command.  I changed the code to make system_command call vanilla
system(cmd), and the display corruption problem went away for _both_ slang
and pdcurses.

Sorry, no diffs, unless someone has the diff program ported to Windoze.