Re: Deep expression chain performance

From	gah4 <gah4@u.washington.edu>
Newsgroups	comp.compilers
Subject	Re: Deep expression chain performance
Date	2023-04-26 13:06 -0700
Organization	Compilers Central
Message-ID	<23-04-014@comp.compilers> (permalink)
References	<23-04-012@comp.compilers>

Show all headers | View raw

On Tuesday, April 25, 2023 at 9:21:44 AM UTC-7, Andy wrote:
> I am writing C grammar.
> Grammar speed may be down caused by deep expression chain.
> For example, simple "n=0" has 20 levels:
> assignmentExpression
> conditionalExpression
> logicalOrExpression
> logicalAndExpression
> inclusiveOrExpression
> exclusiveOrExpression
> andExpression
> eqaulityExpression
> relationalExpression
> shiftExpression
> additiveExpression
> multiplicativeExpression
> castExpression
> unaryExpression
> postfixExpression
> postfixExpressionLeft
> atom
> literal
> IntegerConstant
> DeciamlConstant

A not unusual compiler design in the olden (small memory) days, is recursive
descent
for statements, and operator precedence for expressions.  If you do
expressions with
recursive descent, you get, as you note, 20 levels on the stack.  Operator
precedence
avoids all those levels, and works well for expressions for most languages.

Others have explained how parser generators do it.

Otherwise, the design of small-memory compilers is a lost art.

Back to comp.compilers | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

Deep expression chain performance Andy <borucki.andrzej@gmail.com> - 2023-04-25 06:51 -0700
  Re: Deep expression chain performance anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2023-04-26 16:33 +0000
  Re: Deep expression chain performance gah4 <gah4@u.washington.edu> - 2023-04-26 13:06 -0700
  Re: Deep expression chain performance Kaz Kylheku <864-117-4973@kylheku.com> - 2023-04-27 00:28 +0000
    Re: Deep expression chain performance anton@mips.complang.tuwien.ac.at (Anton Ertl) - 2023-04-27 06:17 +0000

csiph-web