[PATCH] gdbserver: start a gitignore

[PATCH] gdbserver: start a gitignore

[Mike Frysinger]

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
---
 gdb/gdbserver/.gitignore |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
 create mode 100644 gdb/gdbserver/.gitignore

diff --git a/gdb/gdbserver/.gitignore b/gdb/gdbserver/.gitignore
new file mode 100644
index 0000000..9c94950
--- /dev/null
+++ b/gdb/gdbserver/.gitignore
@@ -0,0 +1,4 @@
+gdbreplay
+gdbserver
+reg-*.c
+version.c
-- 
1.7.3.1

Re: [PATCH] gdbserver: start a gitignore

[Joel Brobecker]

> diff --git a/gdb/gdbserver/.gitignore b/gdb/gdbserver/.gitignore
> new file mode 100644
> index 0000000..9c94950
> --- /dev/null
> +++ b/gdb/gdbserver/.gitignore
> @@ -0,0 +1,4 @@
> +gdbreplay
> +gdbserver
> +reg-*.c
> +version.c

I'm OK but slightly hesitant to introduce a .gitignore for files that
are compilation artifacts, not development artifacts.  If we start
on that road, the list is possibly quite long.

I believe that we should encourage people to compile GDB using
a build directory that is different from the source directory.
Except I don't know why I think that! :-). I know that I have been
told that this should be the cannonical way of building GNU projects,
and that the in-source build might one day become unsupported.
I just know that I find it more convenient, precisely because
it avoids polluting my source tree with such build artifacts.

What do people think?

-- 
Joel

Re: [PATCH] gdbserver: start a gitignore

[Yao Qi]

On 12/22/2010 12:45 PM, Joel Brobecker wrote:
>> diff --git a/gdb/gdbserver/.gitignore b/gdb/gdbserver/.gitignore
>> new file mode 100644
>> index 0000000..9c94950
>> --- /dev/null
>> +++ b/gdb/gdbserver/.gitignore
>> @@ -0,0 +1,4 @@
>> +gdbreplay
>> +gdbserver
>> +reg-*.c
>> +version.c
> 
> I'm OK but slightly hesitant to introduce a .gitignore for files that
> are compilation artifacts, not development artifacts.  If we start
> on that road, the list is possibly quite long.
> 
> I believe that we should encourage people to compile GDB using
> a build directory that is different from the source directory.
> Except I don't know why I think that! :-). I know that I have been
> told that this should be the cannonical way of building GNU projects,
> and that the in-source build might one day become unsupported.
> I just know that I find it more convenient, precisely because
> it avoids polluting my source tree with such build artifacts.
> 

Yes, it is not recommended to build in source dir.  The patch looks
good, but not necessary.

-- 
Yao (齐尧)

Re: [PATCH] gdbserver: start a gitignore

[Eli Zaretskii]

> Date: Wed, 22 Dec 2010 08:45:18 +0400
> From: Joel Brobecker <brobecker@adacore.com>
> Cc: gdb-patches@sourceware.org
> 
> I'm OK but slightly hesitant to introduce a .gitignore for files that
> are compilation artifacts, not development artifacts.  If we start
> on that road, the list is possibly quite long.

How long can it be?  Something like *.o, *.so, *.exe, and a couple of
binaries we produce?  Is that really a problem?

I see such exclusions in every GNU project I'm working on.  I don't
think this road is too long ;-)

> I believe that we should encourage people to compile GDB using
> a build directory that is different from the source directory.

Right, but you cannot (and shouldn't, IMO) force people doing that by
such measures.  People who want to build in the source tree will
complain, but will not stop doing that.

> Except I don't know why I think that! :-).

I do.  From gdb/README:

  You can build GDB right in the source directory:

	cd gdb-VERSION
	./configure
	make
	cp gdb/gdb /usr/local/bin/gdb     (or wherever you want)

  However, we recommend that an empty directory be used instead.
  This way you do not clutter your source tree with binary files
  and will be able to create different builds with different 
  configuration options.

Re: [PATCH] gdbserver: start a gitignore

[Joel Brobecker]

> How long can it be?  Something like *.o, *.so, *.exe, and a couple of
> binaries we produce?  Is that really a problem?

I'm not really objecting, nor am I thinking that the list of overly
long. If others are happy with that, that's fine with me too. Note
that you'll also want to add all the files generated during the
configure and build: config.xxx, Makefile, etc.

-- 
Joel

Re: [PATCH] gdbserver: start a gitignore

[Mike Frysinger]

[-- Attachment #1: Type: Text/Plain, Size: 668 bytes --]

On Wednesday, December 22, 2010 06:24:18 Eli Zaretskii wrote:
> > Date: Wed, 22 Dec 2010 08:45:18 +0400
> > From: Joel Brobecker <brobecker@adacore.com>
> > Cc: gdb-patches@sourceware.org
> > 
> > I'm OK but slightly hesitant to introduce a .gitignore for files that
> > are compilation artifacts, not development artifacts.  If we start
> > on that road, the list is possibly quite long.
> 
> How long can it be?  Something like *.o, *.so, *.exe, and a couple of
> binaries we produce?  Is that really a problem?

the higher level gitignores already take care of the common ignores.  subdirs 
need to only mention their specific generated files.
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] gdbserver: start a gitignore

[Joel Brobecker]

> the higher level gitignores already take care of the common ignores.  subdirs 
> need to only mention their specific generated files.

OK for both .gitignore files.

I was nit-picking on myself whether to have one gitignore file for
the entire gdb project, or multiple ones.  I like one because all
the information is grouped in one single file.  On the other hand,
this has several drawbacks: merging conflicts are more likely, and
renaming a directory inside GDB means that the global gitignore
would probably need to be updated as well.  Neither drawback is
significant, but since we have patches that avoid them both...

-- 
Joel

Re: [PATCH] gdbserver: start a gitignore

[Mike Frysinger]

[-- Attachment #1: Type: Text/Plain, Size: 286 bytes --]

On Wednesday, December 22, 2010 23:09:52 Joel Brobecker wrote:
> > the higher level gitignores already take care of the common ignores. 
> > subdirs need to only mention their specific generated files.
> 
> OK for both .gitignore files.

thanks ... ive committed them now
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] gdbserver: start a gitignore

[Michael Snyder]

Eli Zaretskii wrote:
>> Date: Wed, 22 Dec 2010 08:45:18 +0400
>> From: Joel Brobecker <brobecker@adacore.com>
>> Cc: gdb-patches@sourceware.org
>>
>> I'm OK but slightly hesitant to introduce a .gitignore for files that
>> are compilation artifacts, not development artifacts.  If we start
>> on that road, the list is possibly quite long.
> 
> How long can it be?  Something like *.o, *.so, *.exe, and a couple of
> binaries we produce?  Is that really a problem?

*.a, *.la, *.lo....

Re: [PATCH] gdbserver: start a gitignore

[Mike Frysinger]

[-- Attachment #1: Type: Text/Plain, Size: 703 bytes --]

On Wednesday, December 22, 2010 13:44:55 Michael Snyder wrote:
> Eli Zaretskii wrote:
> >> Date: Wed, 22 Dec 2010 08:45:18 +0400
> >> From: Joel Brobecker <brobecker@adacore.com>
> >> Cc: gdb-patches@sourceware.org
> >> 
> >> I'm OK but slightly hesitant to introduce a .gitignore for files that
> >> are compilation artifacts, not development artifacts.  If we start
> >> on that road, the list is possibly quite long.
> > 
> > How long can it be?  Something like *.o, *.so, *.exe, and a couple of
> > binaries we produce?  Is that really a problem?
> 
> *.a, *.la, *.lo....

... which are already handled in the top level and thus not relevant to gdb-
specific gitignore files
-mike

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [PATCH] gdbserver: start a gitignore

[Tom Tromey]

>>>>> "Joel" == Joel Brobecker <brobecker@adacore.com> writes:

Joel> I believe that we should encourage people to compile GDB using
Joel> a build directory that is different from the source directory.
Joel> Except I don't know why I think that! :-).

For a long time, GCC didn't support in-tree builds.  I think this, plus
the convenience of being able to do multiple builds from a single source
tree, or clean a build by "rm -rf", contributed to its adoption.

Joel> What do people think?

I don't mind .gitignore files listing build artifacts.
It is unlikely to break anything and I am not likely to notice anyhow :-)

Tom