[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

lynx-dev pre.10 : -post_data, nls-fork, various

From: Serge MUNHOVEN
Subject: lynx-dev pre.10 : -post_data, nls-fork, various
Date: Wed, 14 Oct 1998 18:08:56 +0200

Dear all

Yesterday I compiled lynx2-8-1pre9 and had a new look at some problems
experienced with pre6 (and a new one ? Sorry wiped out pre6 before noticing
the problem).

- Broken def7_uni.h : thanks to Leonid Pauzner, Thomas E. Dickey and
Michael Warner for their replies. I seem to be the only one affected and did
not yet find the true reason for this problem. My system is the same as
Michael's and has been reinstalled from scratch a few months ago. The problem
still appears in pre9 (no surprise) and also on another system 4.0c. Under
some mysterious circumstances
        printf("%3d", unicount[i]);
is not executed, that's all I can say for now. I hacked around it, but as I
said, I did not find the real answer.

- Now another one that kept me up last night ... (hope somebody will be able
to reproduce it this time). First, the config.status:
# ./configure  --with-zlib --enable-forms-options --enable-persistent-cookies 
--enable-partial --enable-nsl-fork --enable-dired

% cat POST
% cat POST | ./lynx -cfg ./lynx.cfg -post_data > FILE

lynx: Can't access startfile
% cat POST | ./lynx -cfg ./lynx.cfg -post_data > 
% ./lynx -cfg ./lynx.cfg -post_data > FILE

The first one fails, the second and third work. After tracing those and
trying other things (cat POST - | ./lynx ...; works and while cat is waiting
for some input on stdin, lynx ... has already *completed* without failure !)
and wondering if there was a timing problem or a spurious reading from tty I
came to the conclusion that nls-fork is involved :
# ./configure  --with-zlib --enable-forms-options --enable-persistent-cookies 
--enable-partial --enable-dired
and all combinations work. I crosschecked this on another machine. It would be
nice if someone else could try to reproduce this (or wake me up ?).

- Accessing the first time (no cookie set for
them) and submitting the form proposes a cookie. Even if accepting it, the
submission fails. Exit lynx and restart it (with those cookies still in the
jar) and everything works. Note that I have brian j. pardy's cookie-patch
applied. Is there any difference between the handling of fresh cookies and
those from the jar ?

- Suggestion: if I reply "neVer" to a proposed cookie, shouldn't that be
never ever, i.e. be saved in the cookie-file as well ? Probably the most
satisfying would be to have the choice between "no" (once) - "never" (in this
session) - "NEVER" (ever; persistent).

- Cosmetic: when the contents of an <A></A> element happens cross a linebreak,
only the part on the first line is highlighted, except if you came across it
with the cursor (except if the screen is refreshed later on ...). Nothing
serious, just a bit annoying.

As far as time permits, I'll continue investigating all of those (and a few
other minor aspects).


 - Serge

 Serge Munhoven                              Internet: address@hidden

reply via email to

[Prev in Thread] Current Thread [Next in Thread]