lynx-dev
[Top][All Lists]
Advanced

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

Re: lynx-dev PATCH [dev21]: source caching, take 2


From: Leonid Pauzner
Subject: Re: lynx-dev PATCH [dev21]: source caching, take 2
Date: Mon, 12 Apr 1999 11:34:31 +0400 (MSD)

11-Apr-99 23:25 Scott Bigham wrote:
> On Sun, 11 Apr 1999, Scott Bigham wrote:

>> On Sun, 11 Apr 1999, Leonid Pauzner wrote:

>> > I prefer to have a choice between caching sources in temp files or keeping
>> > them in RAM by managing a dynamic list of Cached Stream's or so.

>> Yeah, well, I'm not that good.

> You had to challenge me, didn't you? ;)  Okay, after some pfutzing
> around with HTChunks, I think I've got a working version of memory
> caching of source.  The SOURCE_CACHE option now has settings FILE,
> MEMORY and NONE, with the obvious meanings, defaulting to NONE.  This
> patch replaces my previous patch, so you'll have to back that out first.

Oh, thanks! It also fix cleanup_files problem. Really nice.

Just two minor problems found with my particular environment:

1) when SOURCE_CACHE:MEMORY and we are running on very slow machine
   (i386 in my case) and the document rather long
   the fetching of "\" source may be really slow,
   but no "partial mode" happens: no read-progress counter
   nor redisplay the page while in progress.
   This is done in HTDisplayPartial() and called from HTCopy() and HTFileCopy(),
   so OK when running with SOURCE_CACHE:FILE

2) when SOURCE_CACHE:FILE and I run into the document with DOS/WINDOWS
   end-of-line (cr/lf) style, the cached source has the same line endings
   (at least with DJGPP when html files opened in a text mode)
   and when I press "\" I saw the source with additional blank lines
   between the actual source lines (in a nornal case lynx handle this
   on an early stage). Seems due to opening temp files via LYUtils.c
   functions. Don't know a workaround yet, this seems affect only
   DOS/WINDOWS platforms and can be done laterly.

3) (for Tom) I think #define SOURCE_CACHE should be enabled by default
   both for configure-based platforms and in the DOS src/makefile.* files
   since this will allow more testing (and rather stable already)
   and the default flag in lynx.cfg is NONE.

>                                               -sbigham





reply via email to

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