Re: [htags feature request] generate hypertext of source codedynamically from CVS repository

[Hideki IWAMOTO]

On Sun, 11 May 2003 13:00:00 +0900, Shigio Yamaguchi wrote...
> I think function 1. brings big merit.
> But about function 2., I don't know whether or not it is a big improvement
> when thinking some side effects:
> 
> o We cannot use grep and id-utils from CGI form when realistically thinking.
> o Users depends on CVS repository server. If server is down then we cannot
>   extract source files. If network is slow then we must wait for reply
>   for a long time.
> 
> Furthermore, anyway, we need original source files to generate tag files
> (GTAGS, ...).
> 
> Should we implement function 2?

Function 2 can be added easily later. Let's focus on function 1.

> linux-2.5.9
> ===========
> 
> FILE TYPE                     DISK USAGE
> --------------------------------------------------
> original source files          147 M bytes (  7%)
> tag files (GTAGS,GRTAGS,...)   409 M bytes ( 19%)
> hypertext(for indexes)                 774 M bytes ( 36%)
> hypertext(for source code)     834 M bytes ( 38%)
> ---------------------------------------------------
> Total                         2164 M bytes (100%)
> 
> From above result, we can say like follows:
> 
>       Function 1. reduces 38% of disk space.
>       Function 2. reduces 7% of disk space.
> 

I think that the effect about a disk space of function1 is a little smaller.
However, it is useful at the point of also reducing the execution time of htags.

* Since source files is available, size of tag files can be made small by using 
compact format.
* Hypertext in D/, R/, and Y/ is omissible by generating dynamically using CGI.
  This makes small both of disk space and execution time of htags.
* Hypertext of source code can be made small by compressing by gzip. 
  This saves disk space, but increases the execution time of htags. 


----
Hideki IWAMOTO  address@hidden