compilation process

R

Roman Ziak

Hello,

I posted this message to comp.compilers couple days ago, but it did
not apear there for some reason.

---

I have had this idea about the compilation process, but I may be not
sound enough yet to see the whole picture.

Let's say that instead of classic front-end -> back-end -> link, one would
have:

1. front-end

2. syntax tree linker

3. syntax tree exhaustive optimizer

4. back-end

The difference is that linking process happens on ASTs level rather than
relocatable machine code.

I would just like to hear an opinion from compiler writers if this idea is
worth pursuing.

Thank for all the replies

Roman
 
R

Rob van der Leek

Hello,

I posted this message to comp.compilers couple days ago, but it did
not apear there for some reason.

comp.compilers is a moderated newsgroup, please see the FAQ
(http://www.faqs.org/faqs/compilers/faq/) how to post messages to this
group and what to do when your message does not appear at all.

Your question is off-topic for c.l.c.

Regards,
 
F

Francois Grieu

"Roman Ziak" propose:
Let's say that instead of classic front-end -> back-end -> link,
one would have:

1. front-end
2. syntax tree linker
3. syntax tree exhaustive optimizer
4. back-end

The difference is that linking process happens on ASTs level rather
than relocatable machine code.

I would just like to hear an opinion from compiler writers if this
idea is worth pursuing.

Yes this would work, and enable many optimizations not possible
in more traditional flow. Just integrate steps 3-4-5 in the new
linker, and this is even mostly transparent to the make process,
except for link time, and the need for new libraries. Serious
problem with an interactive debug process, though.

I'd buy a system that can do this. I bet this has/is being tried,
but could not find a product doing it.


François Grieu
 
A

Arthur J. O'Dwyer

comp.compilers is a moderated newsgroup, please see the FAQ
(http://www.faqs.org/faqs/compilers/faq/) how to post messages to this
group and what to do when your message does not appear at all.

Your question is off-topic for c.l.c.

...But would be topical on the unmoderated groups comp.programming
and comp.lang.misc, for example. The latter especially has lots of
people interested in "non-traditional" translation. :)

HTH,
-Arthur
 
R

Roman Ziak

Rob van der Leek said:
comp.compilers is a moderated newsgroup, please see the FAQ
(http://www.faqs.org/faqs/compilers/faq/) how to post messages to this
group and what to do when your message does not appear at all.

Thanks for the link to FAQ. I think my message has been discarded
because I do not have a valid reply e-mail address.

And hence my conflict ... I would like to post to comp.compilers, but
I do not want to be receiving hundreeds e-mails from unscrupulous
spammers.

Roman
 
K

Keith Thompson

Roman Ziak said:
Rob van der Leek said:
Thanks for the link to FAQ. I think my message has been discarded
because I do not have a valid reply e-mail address.

And hence my conflict ... I would like to post to comp.compilers, but
I do not want to be receiving hundreeds e-mails from unscrupulous
spammers.

The comp.compilers FAQ tells you how to contact the moderator.
He can probably help you; we can't.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,774
Messages
2,569,599
Members
45,162
Latest member
GertrudeMa
Top