codehelp / qof Goto Github PK

Hi Neil,

It's been a very long time ... hey I am trying to build the "gnome time tracker" https://github.com/linas/gnotime which has QOF as a dependency. Scouring the interwebs, it appears that the latest version of QOF is version 0.8.4 but the latest that I have (and the latest in this repo) is 0.7.5 Now, that latest version is supposedly on http://qof.alioth.debian.org/ but Alioth was decommissioned in 2019, and moved to "Salsa" which does not have QOF on it. Care to make any comments about this?

p.s. I might make a pull request containing minor updates. Like renaming README to README.md and anything else that pops up.

p.p.s I'm at [email protected] for a more private channel, if anything private is to be said.

Hi! Is this the new official home of the qof project? It used to be on SourceForge. (This is the link I still turned up when searching for the project.) Then it moved to Alioth. Then Alioth was decommissioned. The last archived version of the page on Alioth says the project is obsolete. Nowhere is the existence of this GitHub repository mentioned. It took me awhile to find it, and I'm glad I finally did, because the commits it contains post the 0.8.8 release solved the problem I was having.

If this is the new home for qof, please edit the SourceForge project to refer people here, rather than to the defunct Alioth page. Thanks!

qof (0.8.8 plus 27ffca9) fails to build (with the version of clang included with Xcode 12.4 on macOS 10.15.7):

lexer.c:1955:40: error: implicit declaration of function 'isatty' is invalid in C99 [-Werror,-Wimplicit-function-declaration]
        b->yy_is_interactive = file ? (isatty( fileno(file) ) > 0) : 0;
                                       ^

As of Xcode 12, implicit declaration of functions is an error. If you use a compiler that doesn't do this, you can experience it by adding -Werror=implicit-function-declaration to CFLAGS or just look for the implicit function declaration warning.

Ordinarily this kind of error would have a straightforward fix: include the header that declares the function:

--- lib/libsql/lexer.c.orig	2014-09-23 02:40:17.000000000 -0500
+++ lib/libsql/lexer.c	2022-03-23 17:58:48.000000000 -0500
@@ -40,6 +40,7 @@
 #include <string.h>
 #include <errno.h>
 #include <stdlib.h>
+#include <unistd.h>
 
 /* end standard C headers. */
 

And if I do this, it builds fine for me.

However I'm uncertain in this case because later in lexer.c it has these lines:

#include "sql_parser.h"

/* can conflict with L_SET defined in /usr/include/unistd.h */
#ifdef L_SET
#undef L_SET
#endif

#ifndef YY_NO_UNISTD_H
/* Special case for "unistd.h", since it is non-ANSI. We include it way
 * down here because we want the user's section 1 to have been scanned first.
 * The user has a chance to override it with an option.
 */
#include <unistd.h>
#endif

And sql_parser.h contains the line:

#define YY_NO_UNISTD_H

So the existing #include <unistd.h> in lexer.c never happens.

I don't really understand the concerns being raised in the comments above and don't know how to ascertain whether including unistd.h up top like I've proposed causes any problems. The test suite has the same number of failures (1) regardless of whether I 1. include unistd.h and build normally or 2. don't include unistd.h but suppress the problem by adding -Wno-error=implicit-function-declaration to CFLAGS.