[Libre-soc-bugs] [Bug 422] Migrate away from nmigen Record
bugzilla-daemon at libre-soc.org
bugzilla-daemon at libre-soc.org
Thu Jul 9 12:10:45 BST 2020
Luke Kenneth Casson Leighton <lkcl at lkcl.net> changed:
What |Removed |Added
--- Comment #8 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to whitequark from comment #7)
> So what I do is I assume that the project leadership knows what they're
> doing and I'm probably missing something important, and first ensure that I
> got the details right by asking. Something like:
ha. that's funny. this is exactly what i advise people regarding google
searches. "assume that what you want to know can be found". except the
irony is that i treat it as an "object" - no people involved - despite
the fact that there were clearly people who wrote the google search engine
> > Can you tell me the rationale and roadmap for deprecating Record, as well as the proposed migration path?
> That's it. One sentence. Once you know you got the facts right, feel free to
> express any significant concerns you have, how disruptive it would (or would
> not) be, and so on. Just... don't jump to conclusions, especially when you
> admit it yourself that you had no reason to do so.
so i think we have identified the issue.
* i looked at the bugreport
* it said "Record is being deprecated"
* i saw no indication of a migration plan
* ask questions
* go "argh Record is going away because that's what the bugreport says is
now that we've identified the source of the discrepancy, can i make the
suggestion that, for *other people* who may see that same bugreport, so
that they do not take up your time with the exact same mistake that i
... or, worse, *not use nmigen at all and not contact you because they
assume that Record is being removed and there is absolutely no replacement*...
... to add a very short sentence, "plan and advice needed to be written here
to help people migrate from Record in existing designs"
something to that effect
> If you're frustrated? Find someone to vent on how much of a disruptive idiot
> I am,
except i don't view you as being a "disruptive idiot". now, it may be
the case that you *think* i think that - i simply don't. and for that
i can only apologise that it looked that way.
> Aside: I don't know how are we having this conversation at all. I've never
> heard of an ASIC project three months before tapeout that did not freeze, at
> least, the toolchain completely, much less one upgrading the toolchain to a
> new major version. By all accounts, nothing I can possibly do ought to be
> able to affect your tapeout schedule at all, even in principle.
i know. it's completely mental. we lost several months (a "friend"
needed my help and it cost me, personally, around USD $7k - around 1/2
my yearly income) i won't go into details.
it is what it is.
the idea of freezing the toolchain is a good one (inexperience, there)
although the possible burden of having to maintain a hard fork if bugs
are found is not something i want to add to the list unless absolutely
necessary, and this possibility i view as a minus.
when we have an actual completed design then yes we'll freeze all tools.
i think we've diverged sufficiently from the topic to warrant closing
this bugreport and re-opening a new one.
You are receiving this mail because:
You are on the CC list for the bug.
More information about the libre-soc-bugs