On Fri, 18 Feb 2000 [EMAIL PROTECTED] wrote:
> If you look in LYNews.c, you'll find that there is a test to determine
> whether the external editor generated a nonempty text file whose content
> is to be posted. Unfortunately, at least under Windows of all flavours,
> the test occurs right after the editor is spawned, under the assumption
> that the editor has completed generating its output file.
>
> This is the wrong assumption. Launching the editor under Windows returns
> you immediately to the calling program. Rather than submit a patch, I am
It would be interesting to know how Gygwin-lynx works. I expect it really
waits for editor termination. But there could be problems: if another instance
of the editor is already running, that already running instance of the editor
could open that file for editing (I think so - at least this happens when
typing 'start foo.doc' when word is already running). The question is: what
will happen with 'second' session of the editor? How cygwin behaves in this
case?
May be asking yes/no (as you suggested) in the case of Windows would be
better choice then - to work around of these potential problems.
> pointing this out so that developers can choose whether to force the
> program to wait for the editor to complete before testing the contents
> of the news-post file or simply wait for the user to indicate "yes" or
> "no" to the status-line prompt that follows.
>
> I have naturally made my own choice, and I am posting this from the latest
> February version of lynx.
>
> Oh yes. I had written previously that lynx converts 'v' to '/' on the
> status line text input. This happens when lynx.exe is linked to the latest
> 1.4.1 version of libslang.a. If you link instead to your archived 1.2
> version of libslang, the problem disappears.
Do you mean that pressing 'g' and typing www.yahoo.comvindex.html will be
equal to pressing 'g' and typing www.yahoo.com/index.html ?
[...]
Best regards,
-Vlad
