lynx-dev
[Top][All Lists]
Advanced

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

Re: lynx-dev Re: pre4 (patch) SOURCE CACHE & LYK_SOURCE tiny charse


From: Leonid Pauzner
Subject: Re: lynx-dev Re: pre4 (patch) SOURCE CACHE & LYK_SOURCE tiny charset f
Date: Fri, 21 May 1999 13:48:48 +0400 (MSD)

>      * From: Vlad Harchev <address@hidden>

> On Thu, 20 May 1999, Leonid Pauzner wrote:
>> 19-May-99 05:00 Vlad Harchev wrote:

>> >  Does it fix the charset problem with reloading while in psrcview you
>> > described?
>>
>> No. It has nothing with it. (I have no idea how to fix the problem).
>>

>  I'm not familiar with chartrans handling in lynx while seems you do.
Klaus was the original implementor.

> Here is a description of what should be done.

>  While lynx is in psrc mode, SGML.c converts html it recieves to another html.
>  The code that supports psrc mode in SGML.c doesn't parse values of any
nor does source mode without psrc.

> attribute. When user invokes psrc mode, the chartrans tables are already
> initialized (since the file was viewed in "normal" mode) - so there is no need
> in parsing html and getting the value of charset in psrc mode.
>  So, when user reloads document while in psrc mode, all information about the
> document that was asquired while parsing in HTML.c (including charset info) is
> lost. To fix the bug, it's enough to preserve charset info when reloading
> document (copy info about charset to newdoc or something like this).

Currently, when we switch to source mode, charset stored by LYPushAssumed()
in LYK_SOURCE and then poped back by LYPopAssumed() in HTLoadDocument() just
after (*protocol->load),  so the information is already preserved by faking
UCLYhndl_for_unspec. So this information lost when we reload with ^R but
apparently not immediately.
(Not sure, but (HTParentAnchor)HTMainAnchor->Stages and ...->charset
should probably be faked the same way instead, IMHO).

Worse, this broken in some a strange way: according to the trace log,
UCLYhndl_for_unspec restored back *not* via LYPopAssumed() but later via
HTMLSetCharacterHandling() in mainloop. This is beyond me.

>  Can you implement such bug fix?
I am not at my machine right now, so could try something after weekend only.

>  This can take much more time from me since I'm not familiar with chartrans
> handling in lynx.

>  Best regards,
>   -Vlad




reply via email to

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