|Incremental Compilers/Linkers firstname.lastname@example.org (1993-03-19)|
|Re: Incremental Compilers/Linkers email@example.com (1993-03-22)|
|Re: Incremental Compilers/Linkers firstname.lastname@example.org (1993-03-23)|
|Re: Incremental Compilers/Linkers email@example.com (1993-03-25)|
|Re: Incremental Compilers/Linkers firstname.lastname@example.org (1993-03-24)|
|From:||email@example.com (Elan Feingold)|
|Keywords:||linker, question, comment|
|Organization:||Cornell Theory Center|
|Date:||Fri, 19 Mar 1993 18:36:19 GMT|
Does anyone have any information on any compilers that do incremental code
generation? i.e. You change on line of your source code and the compiler
only recompiles that one line. For the development stage, if the inserted
code is longer than the old code, the compiler could simply generate a
jump to the new code which ended in a jump back.
How about papers on the subject? Anyone have any thoughts on this topic?
It would seem to me at first glance that one of the prerequisites for such
a system would be a tightly integrated development system (editor,
compiler, linker) with emphasis on the editor<->compiler relationship. It
just seems to me that most development systems today are incredibly
inefficient. When I have to wait 3 minutes to generate a new binary when
I change one line of my source, this seems to point towards inefficiency
(correct me if I'm wrong)! Techniques such as incremental compiling and
true dynamic linking seem to offer such an increase in performance that
I'm surprised that they are so rare.
| Elan Feingold
| CS/EE Depts.
| Cornell University
| Ithaca NY 14850
[There's been a series of articles in the JCLT by McKeeman and Aki about
an incremental C system they'd been working on at DEC. Also, TOPLAS a few
years back had an interesting article on an incremental linker that got
greatly improved linker turnaround in exchange for a lot of extra space.
Return to the
Search the comp.compilers archives again.