Re: 68hc11/12/s12x/xgate patch

[Fred Cooke <fred dot cooke at gmail dot com>]

Given that a review of the patch passes code quality measures I would
support most of it making it in, however I absolutely do not support
the XGATE assembler hack making it in. There are multiple technical
reasons for this, as being discussed in the other email thread,
however there is an additional reason, principals. Sean Keys, as
mentioned by James in his submission email, has written a clean sheet
gas implementation for the XGATE that does not compromise the existing
HCS12 assembler by stretching it in directions that don't make sense.
You might wonder why two such gas implementations exist when both
parties are clearly aware of each others existence, a fair question.
The answer is simple: At a time when Sean and I were looking to write
XGATE code in our project, and James had begun work on his patches,
he, in a failed attempt to stifle competition with his competing
commercial project, refused to share this work, as contributed to by
hundreds of other members of the public. Although it is legal to do so
under the GPL, it, in my opinion, certainly isn't moral to do so. Sean
spent a lot of time working on these tools due to James not sharing
his free software in a free way, and it would break my heart, if not
his, to see that work go to waste and James' inferior work go in when
James could have collaborated in an open way from the start. What this
boils down to is that one of these two assemblers is going to be
rendered unused and for the above reasons, along with a slew of
technical reasons, I believe that James' XGATE assembler should not
make it upstream. I've written this as Sean is a very nice guy and at
times too gentle and I'll be damned if I'm going to see his work go to
waste while James' stuff gets used. I apologise for the political
content in this email, however if I'd not sent it, I'd be a poor
friend to Sean and would regret it forever. Someone needs to stick
their neck out and stand up for what's right, it looks like it's my
turn today. In closing, I'd like to thank James for his fixes to other
aspects of the tools, as I, at the least, appreciate that work,
despite his questionable behaviour in the past. Thank you for reading.

Regards,

Fred.

On Thu, Jan 5, 2012 at 11:02 PM, James Murray <jsm@jsm-net.demon.co.uk> wrote:
>
> Find attached my latest revision of a patch to the m68hc11 port
>
> -include somes work from Stephane Carrez that never made it
>
> -Adds support for the S12X derivative of the HC12/S12 family
> (new opcodes added, extended modes, testsuites)
>
> -Adds support for the built-in XGATE co-processor (only exists paired
> with S12X, not in isolation.)
> (gas, bfd, ld, opcodes, testsuite)
>
> -Fix up hc11/12 far call trampoline generation (broken in 2006)
>
> -Change objdump output to use hex numbering with 0x prefixes in a more
> common manner. (Was a confusing mix of hex and decimal with and without
> base prefixes before.)
>
> -Adjust hc11/12 testsuites to reflect that changed format and other
> binutils changes since 2005.
>
> -Enable testsuites for main m68hc11 target. (Many were enabled for m6811
> instead for unknown reasons.)
>
> The core of this work has been in production use (against 2.18) for a
> number of years. Thanks to Sean Keys for giving me feedback on the XGATE
> assembler over the last few weeks.
>
> regards
>
> James Murray