language choice of alfs

Hui Zhou zhouhui at wam.umd.edu
Wed Dec 22 13:47:24 PST 2004


On Wed, Dec 22, 2004 at 03:21:12PM -0500, Jeremy Huntwork wrote:
>Ok. So the idea of the language for alfs is not really up in the air. 
>We've already begun a POC in C, and we can begin calling that base code 
>alfs.  

IIRC, moongoo appears as a POC that commands can be directly extracted 
from LFS book. As the behavior goes, I couldn't tell the difference 
between moongoo and lfscmd. When I asked this question on the list, 
Boris Buegling replies that the difference (between moongoo and 
lfscmd) is moongoo parses the command into usable structures and it 
supports 4 syntaxes and it has the new paralell thingy. My further 
queries were ignored. 

The usable structures I assume is in alfs.h, not much beyond lfscmd 
offers at that point. 4 syntaxes is lfsbook, nalfs tools, and what 
else? Any one would like to comment on the syntaxes because I have 
been showing strong interests on it for some time now. And what is the 
paralell thingy? Is it just -j option to make? 

So moongoo as it stands is just a profile parser. There is no official 
profiles yet unless the default is still nALFS profile (Any final 
say?) nALFS contains quite some structures and hierarchies, how are 
they treated in moongoo? Someone please say something on the official 
profile so we can further talk about how much code base moongoo 
provides. 

I suspect moongoo probably was coded up in a few hours 
during a night, as proof of one concept, not the whole concept. 

>All things considered, it still seems to me that C is the best 
>language for the official backend.  Our two core developers are 
>comfortable with that, 

No argument and quite OK with it too.

>and it drops nicely onto a freshly built LFS system.  

Moongoo at least still need libxml2. nALFS need libcurl too.

>As far as the front-end goes, we'll see when we come to that. 
>As James pointed out the SRS is being written so that it is not language 
>specific.  At this point, any further alfs discussion should really be 
>based on what is already documented on the wiki.  There are still some 
>holes in the SRS, though, so we need to look at those sections in the 
>SRS and offer comments on the missing sections.

If you guys will bring the SRS one by one to this list, I will happily 
join the discussion, and contributing to the holes if that's good.

Regards,

-- 
Hui Zhou



More information about the alfs-discuss mailing list