Re: Why don't compiler writers adhere to the dragon book recommendation of one lexer rule for keywords and identifiers?

Kaz Kylheku <480-992-1380@kylheku.com>
Sun, 26 Jun 2022 06:17:49 -0000 (UTC)

          From comp.compilers

Related articles
| List of all articles for this month |
From: Kaz Kylheku <480-992-1380@kylheku.com>
Newsgroups: comp.compilers
Date: Sun, 26 Jun 2022 06:17:49 -0000 (UTC)
Organization: A noiseless patient Spider
References: 22-06-075 22-06-079
Injection-Info: gal.iecc.com; posting-host="news.iecc.com:2001:470:1f07:1126:0:676f:7373:6970"; logging-data="46227"; mail-complaints-to="abuse@iecc.com"
Keywords: storage, history, comment
Posted-Date: 26 Jun 2022 13:44:13 EDT

On 2022-06-25, gah4 <gah4@u.washington.edu> wrote:
> Some years ago, the Hercules group was trying to get gcc running
> on an emulated IBM S/370 running MVS, with an 8M region.
> (Out of the 16M byte address space, MVS takes up about half.)
> But you can't run gcc in 8M bytes.
>
> When I remember S/370 and OS/VS2, the usual region was 300K,
> which we thought was big.
>
> And now, we can barely run a system with 4G main memory,
> such as the Macbook Air that I am writing this on.


My TXR Lisp hits a peak memory footprint of around 17 megabytes during
the build of its compiler and standard library (on a 32 bit GNU/Linux
system). If the CONFIG_SMALL_MEM build time option is used, it can get
down to 11. That's a total footprint including all the code and data
areas, such as mappings for the C library (which is a pig nowadays), as
reported by common tools like top. (CONFIG_SMALL_MEM just adjusts the
sizes of some static arrays used by the garbage collector and sets
certain thresholds differently.)


--
TXR Programming Language: http://nongnu.org/txr
Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
[I think we should stop here before I start telling you what we did
on a 4K PDP-8. -John]


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.