Re: merging 1.11 with 1.10h9

[Juergen Lock <nox@jelal.north.de>]

In article <9408241605.AA08959@irs.inf.tu-dresden.de> you write:
>Hi!
>
>I'm in the process of merging all the MiNT diffs and versions.  I
>started with merging 1.10h9 and 1.11beta into a single source tree
>(with the help of CVS, this went pretty smoothly).
>
>Appended after my .sig are the relevant differences between 1.11beta
>and the current source tree (patches to 1.11 have not yet been
>applied).  As I'm not as a big MiNT hacker as some of you, I'm asking
>you to compare those differences and help me decide whether they are
>tolerable.

 ok some comments for what might still be unclear... :)

>diff -cr --exclude=*~ --exclude=.#* --exclude-from=merger/mint/.cvsignore --show-c-function 1.11beta/mint/biosfs.c merger/mint/biosfs.c
>*** 1.11beta/mint/biosfs.c	Tue Aug  9 13:33:26 1994
>--- merger/mint/biosfs.c	Wed Aug 24 16:31:53 1994
>*************** bios_fscntl(dir, name, cmd, arg)
>*** 902,907 ****
>--- 902,910 ----
>  		}
>  		b->drvsize = 0;
>  		b->device = &bios_tdevice;
>+ 		if (b->device->readb)
>+ 			/* device has DEVDRV calls beyond unselect */
>+ 			b->drvsize = offsetof (DEVDRV, readb) + sizeof (long);
>  		b->private = arg;
>  		b->flags = O_TTY;
>  		*b->tty = default_tty;
 this made bios ttys added thru DEV_NEWTTY try bios_readb/writeb...
i think currently it has no effect (new devices should be > max_btty) so
removing is ok.

>...
>diff -cr --exclude=*~ --exclude=.#* --exclude-from=merger/mint/.cvsignore --show-c-function 1.11beta/mint/dosfile.c merger/mint/dosfile.c
>*** 1.11beta/mint/dosfile.c	Tue Aug  9 13:33:32 1994
>--- merger/mint/dosfile.c	Wed Aug 24 16:40:08 1994
>...
>*************** do_open(name, rwmode, attr, x)
>*** 236,241 ****
>--- 242,253 ----
>  		return NULL;
>  	}
>
>+ 	if (tofreed) {
>+ 		tofreed->links = 0;
>+ 		free_region(tofreed);
>+ 		tofreed = 0;
>+ 	}
>+ 
>  /* special code for opening a tty */
>  	if (is_terminal(f)) {
>  		extern struct tty default_tty;	/* in tty.c */
 open() a shared text executable has to free its sticky text region
(or it would keep failing until another process needs the memory...)
that is detected in filesys.c:denyshare where the actual free_region
cannot be called and so i moved it here.

>...
>diff -cr --exclude=*~ --exclude=.#* --exclude-from=merger/mint/.cvsignore --show-c-function 1.11beta/mint/tosfs.c merger/mint/tosfs.c
>*** 1.11beta/mint/tosfs.c	Tue Aug  9 13:33:51 1994
>--- merger/mint/tosfs.c	Wed Aug 24 17:40:06 1994
>*************** around:
>*** 687,692 ****
>--- 687,694 ----
>  		xattr->blksize = clsizb[fc->dev];
>  	
>  	xattr->nblocks = (xattr->size + xattr->blksize - 1) / xattr->blksize;
>+ 	if (!xattr->nblocks && (ti->attr & FA_DIR))
>+ 		xattr->nblocks = 1;	/* no dir takes 0 blocks... */
>  	xattr->mdate = xattr->cdate = xattr->adate = ti->date;
>  	xattr->mtime = xattr->ctime = xattr->atime = ti->time;
>  	xattr->mode = (ti->attr & FA_DIR) ? (S_IFDIR | DEFAULT_DIRMODE) :
 well :)  GEMDOG keeps insisting all its directories take up no disk space,
i only wanted `du' print more realistic numbers...

 cheers
	Juergen