I'd like to suggest integrating the new XDRIVER code into the source
tree.
The current mechanism of having "configure" modify several files
(which are under CVS control) suffers from a couple of problems:
1. Running "cvs diff" includes the differences between the *.new
versions of the files and the CVS versions. This is a bit of a
nuisance if you're doing general tidying (minor changes to many
scattered files), where it's simpler to diff the whole tree than to
figure out exactly which files need to be diff'd.
2. Anyone running "cvs commit" without thinking will overwrite the CVS
versions (fifo?) with whichever version is in their build tree at the
time (probably sockets).
If I hadn't encountered (1) first, I probably would have encountered
(2) instead.
--
Glynn Clements <glynn.clements@virgin.net>
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'
On Mon, Apr 09, 2001 at 08:36:58PM +0100, Glynn Clements wrote:
I'd like to suggest integrating the new XDRIVER code into the source
tree.
The current mechanism of having "configure" modify several files
(which are under CVS control) suffers from a couple of problems:
1. Running "cvs diff" includes the differences between the *.new
versions of the files and the CVS versions. This is a bit of a
nuisance if you're doing general tidying (minor changes to many
scattered files), where it's simpler to diff the whole tree than to
figure out exactly which files need to be diff'd.
2. Anyone running "cvs commit" without thinking will overwrite the CVS
versions (fifo?) with whichever version is in their build tree at the
time (probably sockets).
If I hadn't encountered (1) first, I probably would have encountered
(2) instead.
Hi all,
I agree with Glynn. Perhaps we only keep the fifo version in a
subdirectory along with the established mechanism to activate it
(Bernhard wants to keep fifos for old platforms) and put the
sockets version into the default place. I guess that 99% of users
want sockets.
Any objections?
Regards
Markus
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'
Markus Neteler wrote:
> I'd like to suggest integrating the new XDRIVER code into the source
> tree.
>
> The current mechanism of having "configure" modify several files
> (which are under CVS control) suffers from a couple of problems:
>
> 1. Running "cvs diff" includes the differences between the *.new
> versions of the files and the CVS versions. This is a bit of a
> nuisance if you're doing general tidying (minor changes to many
> scattered files), where it's simpler to diff the whole tree than to
> figure out exactly which files need to be diff'd.
>
> 2. Anyone running "cvs commit" without thinking will overwrite the CVS
> versions (fifo?) with whichever version is in their build tree at the
> time (probably sockets).
>
> If I hadn't encountered (1) first, I probably would have encountered
> (2) instead.
Hi all,
I agree with Glynn. Perhaps we only keep the fifo version in a
subdirectory along with the established mechanism to activate it
(Bernhard wants to keep fifos for old platforms) and put the
sockets version into the default place. I guess that 99% of users
want sockets.
I wasn't necessarily suggesting removing/disabling anything. Rather,
merging the alternate versions of each file into a single file, with
feature-test macros selecting which version was actually used.
OTOH, if only a small fraction of users will want something other than
sockets, it may not be worth the effort.
--
Glynn Clements <glynn.clements@virgin.net>
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'
Dear Markus,
Can you please stop mailing me all the GRASS mails.
I have tried to UNSUBSCRIBE GRASS5 but with no sucsess.
Please HELP me!
Thanks,
Tinus Joubert
tinus@medbel.com
----- Original Message -----
From: "Markus Neteler" <neteler@geog.uni-hannover.de>
To: <grass5@geog.uni-hannover.de>
Sent: Tuesday, April 10, 2001 9:28 AM
Subject: Re: [GRASS5] XDRIVER
On Mon, Apr 09, 2001 at 08:36:58PM +0100, Glynn Clements wrote:
>
> I'd like to suggest integrating the new XDRIVER code into the source
> tree.
>
> The current mechanism of having "configure" modify several files
> (which are under CVS control) suffers from a couple of problems:
>
> 1. Running "cvs diff" includes the differences between the *.new
> versions of the files and the CVS versions. This is a bit of a
> nuisance if you're doing general tidying (minor changes to many
> scattered files), where it's simpler to diff the whole tree than to
> figure out exactly which files need to be diff'd.
>
> 2. Anyone running "cvs commit" without thinking will overwrite the CVS
> versions (fifo?) with whichever version is in their build tree at the
> time (probably sockets).
>
> If I hadn't encountered (1) first, I probably would have encountered
> (2) instead.
Hi all,
I agree with Glynn. Perhaps we only keep the fifo version in a
subdirectory along with the established mechanism to activate it
(Bernhard wants to keep fifos for old platforms) and put the
sockets version into the default place. I guess that 99% of users
want sockets.
Any objections?
Regards
Markus
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write
to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'
On Tue, Apr 10, 2001 at 08:47:53AM +0100, Glynn Clements wrote:
Markus Neteler wrote:
> > I'd like to suggest integrating the new XDRIVER code into the source
> > tree.
> >
> > The current mechanism of having "configure" modify several files
> > (which are under CVS control) suffers from a couple of problems:
> >
> > 1. Running "cvs diff" includes the differences between the *.new
> > versions of the files and the CVS versions. This is a bit of a
> > nuisance if you're doing general tidying (minor changes to many
> > scattered files), where it's simpler to diff the whole tree than to
> > figure out exactly which files need to be diff'd.
> >
> > 2. Anyone running "cvs commit" without thinking will overwrite the CVS
> > versions (fifo?) with whichever version is in their build tree at the
> > time (probably sockets).
> >
> > If I hadn't encountered (1) first, I probably would have encountered
> > (2) instead.
>
> Hi all,
>
> I agree with Glynn. Perhaps we only keep the fifo version in a
> subdirectory along with the established mechanism to activate it
> (Bernhard wants to keep fifos for old platforms) and put the
> sockets version into the default place. I guess that 99% of users
> want sockets.
I wasn't necessarily suggesting removing/disabling anything. Rather,
merging the alternate versions of each file into a single file, with
feature-test macros selecting which version was actually used.
OTOH, if only a small fraction of users will want something other than
sockets, it may not be worth the effort.
There are some pretty substantial differences between the drivers. Not
to mention the same would have to be done for the libraster, and the
driverlib used by CELL/HTMLMAP/PNG drivers. Easier just to remove the
cvs versions of the files that get swapped during configure. There's a
copy of each fifo/ipc/socket variant in a separate subdirectory, which
are the only copies that need to be maintained.
--
Eric G. Miller <egm2@jps.net>
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'
Hello all
"Eric G. Miller" wrote:
On Tue, Apr 10, 2001 at 08:47:53AM +0100, Glynn Clements wrote:
>
> Markus Neteler wrote:
> > I agree with Glynn. Perhaps we only keep the fifo version in a
> > subdirectory along with the established mechanism to activate it
> > (Bernhard wants to keep fifos for old platforms) and put the
> > sockets version into the default place. I guess that 99% of users
> > want sockets.
>
> I wasn't necessarily suggesting removing/disabling anything.
> Rather, merging the alternate versions of each file into a single
> file, with feature-test macros selecting which version was actually
> used.
>
> OTOH, if only a small fraction of users will want something other
> than sockets, it may not be worth the effort.
There are some pretty substantial differences between the drivers.
Not to mention the same would have to be done for the libraster, and
the driverlib used by CELL/HTMLMAP/PNG drivers. Easier just to
remove the cvs versions of the files that get swapped during
configure. There's a copy of each fifo/ipc/socket variant in a
separate subdirectory, which are the only copies that need to be
maintained.
This was discussed before and it was agreed that the fifos would remain
for 5.0 but be removed for 5.1 barring any further problems with
sockets. To solve Glynn's problem can we use
cvs update -P
to remove the files before committing? Or does -P only work with
directories?
--
Sincerely,
Jazzman (a.k.a. Justin Hickey) e-mail: jhickey@hpcc.nectec.or.th
High Performance Computing Center
National Electronics and Computer Technology Center (NECTEC)
Bangkok, Thailand
People who think they know everything are very irritating to those
of us who do. ---Anonymous
Jazz and Trek Rule!!!
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo@geog.uni-hannover.de with
subject 'unsubscribe grass5'