Issue 0140: What does performed and other operation mean?

This issue has been automatically converted from the original issue lists and some formatting may not have been preserved.

Authors: Andy Pepperdine, BSI
Date: 1994-07-27
Submitted against: C90
Status: Closed
Converted from: dr.htm, dr_140.html

Subclause 7.9.5.6 says:

The setvbuf function may be used only after the stream ... has been associated with an open file and before any other operation is performed on the stream.

There are two related questions associated with this statement.

1. What does “performed” mean?

a. Does it include attempts that failed (such as fread on output file, etc.)?

b. In particular, does it include a failed attempt to setvbuf?

c. What about fprintf(f, "")?

2. What does “other operation” mean?

a. Does it include setvbuf itself?

b. Are ferror and feof operations?

c. What about clearerr?

Reasons for asking:

It would seem reasonable to try to get a very large buffer in some applications by attempting to do a setvbuf with, say, 1 MB of buffer space. If that fails, try again with 0.5 MB, etc. Is this allowed?

My guess as to the interpretation is as follows:

1. An operation is “performed” even if it fails for whatever reason.

2. All functions defined in subclause 7.9 are to be treated as “operations.”

This is unsatisfactory, as the above approach of attempting to find a good buffer size would fail.

In the Rationale, it states “The general principle is to provide portable code with a means of requesting the most appropriate popular buffering style, but not to require an implementation to support these styles.” [Emphasis added.]

I interpret this as saying that setvbuf is an advisory call and need not be acted on. However, my questions above still stand as there seems to be no way of negotiating an agreement on good acceptable buffer sizes.

I believe that a clarification is required.


Comment from WG14 on 1997-09-23:

Response

As you say, “setvbuf is an advisory call and need not be acted on.” That is to say, the C Standard allows it to fail. Therefore, discussions of detailed constraints such as you describe could only constitute non-normative advice to programmers or implementers. The Committee does not have any specific advice to give in this regard.