[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: fs-test "unopened FS object" tests: fair or unfair?

From: Glenn A. Thompson <gthompson_at_cdr.net>
Date: 2004-04-27 13:40:11 CEST


Greg Hudson wrote:

>fs-test.c has this test where it invokes svn_fs_new() and then tries
>calling various functions on it without calling svn_fs_open() or
>svn_fs_create(). Currently, my loader library seg faults if you try
>this, attempting to call through a null vtable pointer.
>I can add checks in each of the functions which call through an FS
>vtable pointer, but to my mind, this kind of operation is like calling
>fopen(NULL, NULL); the expected response is an assertion failure or
>core dump, not a reportable error. So I'd prefer to just delete that
>test. What are other people's opinions?
My opinion is to get rid of svn_fs_new. Have the open and create do the
svn_fs_new work. I talk about it in my document in my API visible
changes section. I even mention the fact that this test could go away.


To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Apr 27 13:45:31 2004

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.