Moving dgen_ scripts out of validator_arm into a common directory.

Moving dgen_ scripts out of validator_arm into a common directory.

The dgen_* scripts have been lifted into a common directory so other architectures, such as MIPS, can reuse the scripts without duplicating the code.

BUG= http://code.google.com/p/nativeclient/issues/detail?id=2275
TEST=none

[petarj, 2012-04-23 20:53:36]

Per code review at http://codereview.chromium.org/9979025/, we need to move
dgen_* scripts into a common directory, so they can be reused elsewhere.

Looking forward to getting your comments.

Petar

[Nick Bray (chromium), 2012-04-24 00:45:29]

I'd put the scripts inside a "dgen" directory, to keep things tidy.

What parts of these scripts are ARM specific, and what will you need to modify
to support MIPS?  You probably don't need to do anything else in the CL, but I
am trying to visualize what the next step is.

[petarj, 2012-04-24 18:14:25]

The change has been updated. The scripts have been moved into validator/dgen
directory.

@Nick
To answer your question, in previous version, the scripts have not been modified
for MIPS (i.e. they were used as-is). I hope there will be no need for further
changes on these scripts.

[Nick Bray (chromium), 2012-04-24 19:57:56]

Petar, meet Karl.  Karl, meet Petar.

In my zeal to prevent code from being duplicated, I overlooked the fact that
Karl is making a lot of changes to the ARM validator.  This creates two
problems:

1) file move merge conflicts
2) if code is shared, changes to the ARM validator may break MIPS.

From this perspective, duplicating the code would help isolate the two
validators.  On the other hand, MIPS may want many of the improvements being
done for ARM.

So: Petar, meet Karl.  Karl, meet Petar.  You both know the situation better
than I do.  I'd like to see minimal code duplication, but pragmatic concessions
may be needed to avoid workflow issues.  Figure out what you both need?

For Karl, this is the original CL which will includes the ISA table:
http://codereview.chromium.org/9979025/

[petarj, 2012-04-24 22:16:09]

Nick, Karl,

if the changes Karl is introducing are not going to be arch-specific or will be
to a modest extent, and if it will be possible to use the scripts the same way
for MIPS and ARM architecture, we can take the burden of updating scripts/scons
when there is a need for it.

Otherwise, we can leave the copies in CL 9979025, and think about this again
after the changes for ARM are done.

What would you prefer?

[Karl, 2012-04-24 22:32:22]

Petarj,

Right now, I'm working on building a good testing strategy for the decoder
tables, and I am basing it on ARM. As such, it is in a state of flux. My
suggestion is to use a copy. It should be fairly easy to merge the scripts once
I have settled how best to handle the testing requirements I have.

[petarj, 2012-04-24 22:40:46]

Karl,

I am fine with that.

Nick,

are you fine then that we leave the copies in 9979025 as they are, and change
them once Karl sorts out the ARM version?

[Nick Bray (chromium), 2012-04-24 22:48:12]

Sure, although I'd still suggest keeping them in a directory.  Sorry for running
you in a circle.

[Nick Bray (chromium), 2012-04-24 22:51:39]

I should mention, in case you haven't done this yet, when you want me to take
another look at previous CL, just reply "PTAL" (please take another look) on the
review.