| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is long overdue. Let's see if it gets me using traces more during
debugging.
Though perhaps I'm being too persnickety. These are all valid ways to
debug programs:
a) print directly to screen
b) log, and then dump the log on some condition
c) temporarily print selected log statements directly to screen
d) log, and then browse the log using the zoom interface
For a) to work we need to normally keep prints empty.
For b) to work the log needs to be of some manageable size, where it's
tractable to find interesting features.
d) is the ultimate weapon, but might be slow because it's interactive
c) seems like the ugly case. Should I be trying to avoid it altogether?
Let's try, and see if d) is useable when we want to do c). For simple
cases it's still totally acceptable to just print. If the prints get too
complex to parse, then we move to the zoom interface. Hopefully it'll be
easier because we have to spend less time getting the prints just so.
(Independent of all this, often the best way to make a log manageable so
any of the approaches works: distill the bad behavior down to a test.
But that leads to chicken-and-egg situations where you need to first
understand before you can distill.)
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
| |
The idea is that to-text-line should truncate blindly past some
threshold, even if to-text isn't smart enough to avoid infinite loops.
Maybe I should define a 'truncating buffer' which stops once it fills
up. That would be an easy way to eliminate all infinite loops in
to-text-line.
|
| |
|
|
|
|
| |
Drop the display-list test because it's subsumed by the stash-on-list
test right below, which doesn't require the notion of a screen from
later layers.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
| |
Just one tangle of dependencies left in the transforms. Untangling it
will require one more region: for transforms modifying types (inside
instructions).
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
Some more structure to transforms, and flattening of dependencies
between them.
|
| |
|
|
|
|
|
| |
In general you only want to specify one transform in terms of
(before/after) another if the other direction wouldn't work. Otherwise
try to make it work by just pushing transforms at the start/end of the
list.
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Ok, now I'm a little happier about our type checking. Type checking lies
in three layers:
layer 21: checking types, usually at run-time in the VM or just before
layer 57: checking type names
layer 59: checking type names
It's taken me the process of writing this commit to convince myself that
all three layers check mappings for literal values in a consistent
manner. Layer 21 uses sizes, and layers 57/59 have whitelists, but
either way the goal is that number != character != boolean unless
mediated through a literal.
|
| |
|
|
|
| |
More cleanup. Haven't bothered to figure out why the trace for
specialize_with_literal_4 is repeatedly perturbed.
|
| | |
|
| |
|
|
|
| |
Eliminate one of several hacky type-check helpers; they've been
proliferating lately.
|
| |
|
|
|
| |
'append' also implicitly calls 'to-text' unless there's a better
variant.
|
| |
|
|
| |
Literal '0' ingredients should map to numbers before addresses.
|
| | |
|
| |
|
|
|
|
|
| |
A long-standing question resolved: why specializations sometimes skipped
some names. Turns out cleanup is incomplete if Recipe_ordinal and Recipe
aren't exactly lined up with each other, and the early exit in
new_variant was breaking that constraint.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
| |
It's just the other direction we want to avoid.
|
| |
|
|
| |
2473 was the final bugfix holding this back.
|
| |
|
|
|
| |
/raw is to express absolute addresses
/unsafe is to sidestep type-checking in test setup
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
| |
This turned out to be surprisingly easy: rather than try to specify the
types expected by each operation, just have primitive instructions go
through any other variants and only remain unnamed if no variants pass.
All I had to do was stop using contains_key() anywhere on Recipe_variants,
so that I could use get_or_insert() rather than get().
|
| |
|
|
|
|
|
|
|
| |
Since we switched to end() for terminating trace lines, there's a lot
less reason to avoid this. We don't nest trace statements either
anymore.
I'd like to not hide warnings and still be able to make assertions on
their absence so that printed warnings also express as failed tests.
|
| | |
|
| |
|
|
|
|
|
|
| |
Not entirely happy with this. Maybe we'll find a better name. But at
least it's an improvement.
One part I *am* happy with is renaming string-replace to replace,
string-append to append, etc. Overdue, now that we have static dispatch.
|
| | |
|
| |
|
|
| |
Thanks Caleb Couch.
|
| |
|
|
|
| |
Requires carefully deleting specializations so that they can be
reintroduced each time.
|
| | |
|
| | |
|
| | |
|