Re: ACT tree new files

Re: ACT tree new files

[Robert Dewar]

At least for some of these cases there are good reasons for not including
the files, e.g. in some cases these files are relevant only for GNAT 3.
I would recommend holding off uploading any files until we complete the
initial merge, because otherwise things will be inconsistent. When the
merge is complete, we can review additional files that make sense to move.

Re: ACT tree new files

[Laurent Guerby]

On Fri, 2003-04-25 at 15:51, Robert Dewar wrote:
> At least for some of these cases there are good reasons for not including
> the files, e.g. in some cases these files are relevant only for GNAT 3.
> I would recommend holding off uploading any files until we complete the
> initial merge, because otherwise things will be inconsistent. When the
> merge is complete, we can review additional files that make sense to move.

My reasoning behind an early addition is that it makes
later bulk merge patches easier to make since you don't
have to find the set of files to add each time which
will either be time consuming trial and error, human reasoning or
you have to spend a time to write a semi automated tool.

Once everything is in place, we'll then be able to safely
cvs delete GNAT 3 files, they were part of the GNU project
at some point anyway, and someone with GNAT 3.15p sources
will be able to find in the FSF CVS that the file was
purposely obsoleted.

Before sending my original email I did a random
sampling of a dozen files and they were all needed
(from Makefile or with closure) but for gnathtml.pl,
with some statistical confidence there must
not be that many files that we'll need to
delete later on.

Assuming my bootstrap succeeds, ok to add the files?

-- 
Laurent Guerby <guerby@acm.org>

Re: ACT tree new files

[Geert Bosch]

On Friday, Apr 25, 2003, at 15:58 America/New_York, Laurent Guerby 
wrote:
> There's of course no urgency, in answer to my question
> about outside help, Geert proposed to accept prepared
> patches, adding files was looking like an easy and useful
> step for the rest of process.

I agree with you. Adding missing files that will be needed later is
a useful step. Thanks for helping with this.

> I assume that Geert will have to do all the work then.

Oh, no! :)

Re: ACT tree new files

[Robert Dewar]

probably they should have copyright notices, though of course such notices
have absolutely no legal significance whatsoever. I will make sure these
three files get copyright notices.

Re: ACT tree new files

[Laurent Guerby]

On Fri, 2003-04-25 at 23:30, Robert Dewar wrote:
> it's fine to prepare the patch, and perhaps it will be helpful, but Geert
> has not even had a chance to look at this patch, 

The patch has no interest (it's all +), only the list of files has.
I spotted one error on my list (final.c has been renamed
to adafinal.c in the FSF tree), I'll submit a proper
list and ChangeLog to gcc-patches over the week-end after
I have a successful bootstrap.

> since he is away for a
> few days, and all I am asking is please don't rush to apply this patch
> without coordination with geert, since I am worried that it may make
> things harder. 

No problem, I'm just asking for approval there.

One thing I assume is in your court: to look at wether the following
ACT tree files need a copyright notice:

vMakefile.adalib
vMakefile.declib
vms-dwarf2.opt

It's all VMS so out of my testing league :).

-- 
Laurent Guerby <guerby@acm.org>

Re: ACT tree new files

[Robert Dewar]

it's fine to prepare the patch, and perhaps it will be helpful, but Geert
has not even had a chance to look at this patch, since he is away for a
few days, and all I am asking is please don't rush to apply this patch
without coordination with geert, since I am worried that it may make
things harder. Perhaps I am wrong, but there is absolutely no day-to-day
urgency here, so let's wait a few days till Geert can coordinate withy
you on this.

Re: ACT tree new files

[Laurent Guerby]

On Fri, 2003-04-25 at 20:30, Robert Dewar wrote:
> > My reasoning behind an early addition is that it makes
> > later bulk merge patches easier to make since you don't
> > have to find the set of files to add each time which
> > will either be time consuming trial and error, human reasoning or
> > you have to spend a time to write a semi automated tool.
> 
> It will definitely be better to delay adding these files until we have
> completed the current merge operation, since it will just complicate
> that merge. It really is not a good idea to simply copy files for the
> sake of copying them without understanding the reasons why they were
> not there in the first place. 

I must admit I don't understand your point. What the merge will do
is just to add all these files, you cannot merge say scn.ads
without adding scng.* which in turn will require merging
errout.* which will need adding erroutc.*, etc...

One missing in your patchset, and you're back to reboostrap.
There are more than 600 files to merge, missing files
can cost a lot of time.

> It certainly will not make things easier!

...

> Are there any files there where for some reason you think there is a day
> to day urgency to add them?

There's of course no urgency, in answer to my question
about outside help, Geert proposed to accept prepared
patches, adding files was looking like an easy and useful
step for the rest of process.

I assume that Geert will have to do all the work then.

-- 
Laurent Guerby <guerby@acm.org>

Re: ACT tree new files

[Robert Dewar]

> My reasoning behind an early addition is that it makes
> later bulk merge patches easier to make since you don't
> have to find the set of files to add each time which
> will either be time consuming trial and error, human reasoning or
> you have to spend a time to write a semi automated tool.

It will definitely be better to delay adding these files until we have
completed the current merge operation, since it will just complicate
that merge. It really is not a good idea to simply copy files for the
sake of copying them without understanding the reasons why they were
not there in the first place. 

It certainly will not make things easier!

Are there any files there where for some reason you think there is a day
to day urgency to add them?

ACT tree new files

[Laurent Guerby]

I looked at the files present in the ACT tree and not in
the FSF ones. Of those 187 look "interesting" and 32 "uninteresting"
(various diff and packaging helpers), both lists are at the end
of this mail (my estimate :).

Of the "interesting", 3 files have no Copyright notice, 
I assume they must be added before getting into the FSF CVS:

vMakefile.adalib
vMakefile.declib
vms-dwarf2.opt

I assume all the "interesting" files are ok to add to FSF CVS? (But
for the three above pending Copyright notice fix). ChangeLog
will be "... : New files", to follow the convention used
by previous merges (see 2002-03-07).

Of the "uninteresting", one still has an RCS Tag:

gnathtml.pl

May be this one can be added to the FSF tree later, looks
useful to be installed, but not used during build.

ada-tree.def has an RCS Tag in ACT CVS, but not in
ACT gcc-head branch, so not a problem for the merge
since gcc-head takes precedence (but might
be worth fixing on ACT side).

Laurent

Interesting List:

1aexcach.adb
1aexexda.adb
1aexexpr.adb
1aexextr.adb
1aexstat.adb
1ainterr.adb
1areatim.adb
1areatim.ads
1aretide.adb
1asytaco.adb
1asytaco.ads
1atags.adb
1ataside.adb
1sinterr.adb
1sinterr.ads
1sosinte.adb
1staprob.adb
1staprob.ads
1staprop.adb
1staprop.ads
1starest.adb
1staskin.adb
1staskin.ads
1stposen.adb
1stposen.ads
3psoccon.ads
3vexpect.adb
3vsoccon.ads
3vsocthi.adb
3vsocthi.ads
3zsoccon.ads
3zsocthi.adb
3zsocthi.ads
51system.ads
55system.ads
56osinte.adb
56osinte.ads
56taprop.adb
56taspri.ads
57system.ads
58system.ads
59system.ads
5aml-tgt.adb
5csystem.ads
5dsystem.ads
5fosinte.adb
5msystem.ads
5psystem.ads
5rsystem.ads
5sml-tgt.adb
5sosprim.adb
5stpopsp.adb
5tsystem.ads
5usystem.ads
5vml-tgt.adb
5wml-tgt.adb
5xparame.ads
5xsystem.ads
5xvxwork.ads
5yparame.ads
5zparame.ads
5ztaspri.ads
5ztfsetr.adb
5ztpopsp.adb
7stfsetr.adb
7straces.adb
7strafor.adb
7strafor.ads
7stratas.adb
address_operations.adb
a-excach.adb
a-exexda.adb
a-exexpr.adb
a-exextr.adb
a-exstat.adb
bld.adb
bld.ads
bld-io.adb
bld-io.ads
ctrl_c.c
current_exception.ads
erroutc.adb
erroutc.ads
errutil.adb
errutil.ads
err_vars.ads
final.c
g-bubsor.adb
g-bubsor.ads
g-comver.adb
g-comver.ads
g-ctrl_c.ads
generic_fast_elementary_functions.adb
get_task_info.adb
get_task_info.ads
g-excact.adb
g-excact.ads
g-heasor.adb
g-heasor.ads
gnat_wrapper.adb
g-pehage.adb
g-pehage.ads
g-perhas.ads
gpr2make.adb
gpr2make.ads
gprcmd.adb
gprep.adb
gprep.ads
g-string.adb
g-string.ads
i-vthrea.adb
i-vthrea.ads
i-vxwoio.adb
i-vxwoio.ads
Makefile.generic
Makefile.prolog
Makefile.rtl
Makefile.vms
Makefile.vms_gcc
mingw32.h
prep.adb
prep.ads
prepcomp.adb
prepcomp.ads
prj-err.adb
prj-err.ads
s-boarop.ads
s-carsi8.adb
s-carsi8.ads
s-carun8.adb
s-carun8.ads
s-casi16.adb
s-casi16.ads
s-casi32.adb
s-casi32.ads
s-casi64.adb
s-casi64.ads
s-casuti.adb
s-casuti.ads
s-caun16.adb
s-caun16.ads
s-caun32.adb
s-caun32.ads
s-caun64.adb
s-caun64.ads
scng.adb
scng.ads
s-exnint.adb
s-exnllf.adb
s-exnlli.adb
s-expint.adb
s-explli.adb
s-geveop.adb
s-geveop.ads
s-hibaen.ads
s-htable.adb
s-htable.ads
sinput-c.adb
sinput-c.ads
s-memcop.ads
socket.c
s-purexc.ads
s-stopoo.adb
s-strcom.adb
s-strcom.ads
s-strxdr.adb
s-tpae65.adb
s-tpae65.ads
s-tporft.adb
styleg.adb
styleg.ads
styleg-c.adb
styleg-c.ads
s-veboop.adb
s-veboop.ads
s-vector.ads
synchronize_nonreentrant_access.adb
system_runtime_tuning.adb
tb-alvms.c
tb-alvxw.c
tempdir.adb
tempdir.ads
vMakefile.adalib
vMakefile.declib
vms-dwarf2.opt
vmshandler.asm
vxaddr2line.adb

Uninteresting List:

binutils-291.dif
declibbuild.com
declibcompile.com
declibgnat1.diff
declibgnat.diff
gas-281.dif
gcc-32.dif
gdb-4.17.gnat.lynxos.diff
gdb-4.17.gnat.tornado2.diff
gdb-4.17.gnat.vxworks.diff
gdb-4.17.gnat.vxworks.tornadoII.diff
gdb-5.3.gnat.tornado.diff
gdb-5.3.tornado.diff
gdb.hlp
gdbkitinstal5.com
gdbkitinstal.com
gdbtk-4.17.diff
gnat-5_OpenVMS_binutils-291.dif
gnat-5_OpenVMS.dif
gnat-5_Windows_NT.dif
gnat.hlp
gnathtml.pl
kitinstal5.com
kitinstal.com
patch.binutils-2.13.2.1-powerpc-elf
patch.powerpc-elf
patch.powerpc-wrs-vxworks
patch.sparc64-wrs-vxworks
patch.xscale-coff-vxworks
patch.xscale-wrs-vxworks
README.BUILD
vmssymvec.com