Proposal: Roundtrip serialization of Cmm (parser-compatible pretty-printer output)

[Andreas Klebinger]

The idea of making Cmm roundtripable comes up every now and then.
While the ability to feed dump output to GHC for debugging or similar 
purposes is useful In the end we always
ended up prioritizing one of the many other things that needed doing.

Or in other words making Cmm (more) roundtripable seems inherently useful.
However it's questionably how much it is worth breaking things like .cmm 
code that exists in libraries for it.
So if you want to work towards this it should be with the goal to avoid 
breakage.

There are likely also a lot of corner cases to consider. Which might 
make this more complicated then it sounds.
Ultimately this is up to you and your mentor. But if I understand 
correctly you have about 5 weeks left for
GSoC so getting full Cmm roundtrip ability into a state where it can be 
merged into GHC during that time might be
too optimistic depending on your haskell/parser/GHC experience.

As a GHC maintainer for us the most useful thing therefore would be 
incremental patches which take Cmm closer
to being roundtripable. And that would allow you to get at least some 
work that benefits the GHC project into the tree even if you end up not 
making it all the way to full roundtrip capability.

On the pure technical aspects:
-------------

> > Create a separate parser ...

1. Creating a separate parser is not viable. It would likely bitrot and 
break on the next change to Cmm and only causes increased maintenance 
overhead. At least not if you want the GHC team to maintain it.

> Extend the current parser with a dedicated block
Having blocks ala C seems fine. Your suggestion seems different however. 
It's unclear from your example how those blocks would work exactly. Is 
`|low_level_unwrapped` |a label. If so can we goto to it? Is it a 
keyword? Something else entirely?

If the main issue is the "offset" string in the generated case I'm fine 
with deleting that from the pretty printer. I'm not sure that does 
anything of value so removing it from the output seems fine. (See 
pprCmmGraph).

 > If we introduce this new “exact” low-level form, it's possible the 
existing low-level mode could become redundant. We might then have:

What changes are you planning that make the new parser/syntax 
incompatible with the old one? Can't you just modify the current parser, 
maybe with some slight changes to the pretty printer, in a way that 
makes it mostly backwards compatible?

 > |aeson| adds a large dependency footprint, and likely wouldn't be 
suitable for inclusion in GHC.

Yes aeson seems unsuitable.

 > Lastly—I’ve heard that parts of the Cmm pipeline may currently be 
under refactoring.

This is the first time I hear of this so I wonder where this information 
came from? There could always be changes to those sorts of things, 
because at the end of the day they are compiler internals. But I'm not 
aware of any big planned changes in the near future.

Cheers
Andreas

On 28/07/2025 02:16, Diego Antonio Rosario Palomino wrote:
>
> Hello GHC devs,
>
> I'm currently working on Cmm documentation and tooling improvements as 
> part of my Google Summer of Code project. One of my core goals is to 
> make Cmm roundtrip serializable.
>
> Right now, the in-memory Cmm data structure—generated programmatically 
> (e.g., from STG via GHC)—can be pretty-printed, and Cmm can also be 
> parsed. However, the pretty-printed version is not compatible with the 
> parser. That is, we cannot take the output of the pretty printer and 
> feed it directly back into the parser.
>
> Example:
>
> Parseable version:
>
> |sum { cr: bits64 x; x = R1 + R2; R1 = x; jump %ENTRY_CODE(Sp(0))[R1]; } |
>
> Pretty-printed version:
>
> |sum() { // [] { info_tbls: [] stack_info: arg_space: 8 } {offset cf: 
> // global _ce::I64 = R1 + R2; R1 = _ce::I64; call (I64[Sp + 0 * 
> 8])(R1) args: 8, res: 0, upd: 8; } } |
>
> Another example:
>
> Parseable version:
>
> |simple_sum_4 { // [R2, R1] cr: // global bits64 _cq; _cq = R2; bits64 
> _cp; _cp = R1; R1 = _cq + _cp; jump (bits64[Sp])[R1]; } |
>
> Pretty-printed version:
>
> |simple_sum_4() { // [] { info_tbls: [] stack_info: arg_space: 8 } 
> {offset cs: // global _cq::I64 = R2; _cr::I64 = R1; R1 = _cq::I64 + 
> _cr::I64; call (I64[Sp])(R1) args: 8, res: 0, upd: 8; } } |
>
> While it’s possible to write parseable Cmm that resembles the 
> pretty-printed version (and hence the internal ADT), they don’t fully 
> match—mainly because the parser inserts inferred fields using 
> convenience functions.
>
> Proposal:
>
> To make roundtrip serialization possible, I propose supporting a new 
> syntax that matches the pretty printer output exactly.
>
> There are a couple of design options:
>
> 1.
>
>     Create a separate parser that accepts the pretty-printed syntax.
>     Files could then use either the current parser or the new strict one.
>
> 2.
>
>     Extend the current parser with a dedicated block syntax like:
>
> |low_level_unwrapped { ... } |
>
> This second option is the one my mentor recommends, as it may better 
> reflect GHC developers' preferences. In this mode, the parser would 
> not insert any inferred data and would expect the input to match the 
> pretty-printed form exactly.
>
> This would enable a true roundtrip:
>
>  *
>
>     Compile Haskell to Cmm (in-memory AST)
>
>  *
>
>     Pretty-print and write it to disk (wrapped in low_level_unwrapped
>     { ... })
>
>  *
>
>     Later read it back using the parser and continue with codegen
>
> Optional future direction:
>
> As a side note: currently the parser has both a “high-level” and a 
> “low-level” mode. The low-level mode resembles the AST more closely 
> but still inserts some inferred data.
>
> If we introduce this new “exact” low-level form, it's possible the 
> existing low-level mode could become redundant. We might then have:
>
>  *
>
>     High-level syntax
>
>  *
>
>     New low-level (exact)
>
>  *
>
>     And possibly deprecate the current low-level variant
>
> I’d be interested in your thoughts on whether that direction makes sense.
>
> Serialization libraries?
>
> One technically possible—but likely unacceptable—alternative would be 
> to derive serialization via a library like |aeson|. That would enable 
> serializing and deserializing the Cmm AST directly. However, I 
> understand that |aeson| adds a large dependency footprint, and likely 
> wouldn't be suitable for inclusion in GHC.
>
> Final question:
>
> Lastly—I’ve heard that parts of the Cmm pipeline may currently be 
> under refactoring. If that’s the case, could you point me to which 
> parts (parser, pretty printer, internal representation, etc.) are 
> being modified? I’d like to align my efforts accordingly and avoid 
> conflicts.
>
> Thanks very much for your time and input! I'm happy to iterate on this 
> based on your feedback.
>
> Best regards,
> Diego Antonio Rosario Palomino
> GSoC 2025 – Cmm Documentation & Tooling
>
>
> _______________________________________________
> ghc-devs mailing list
> ghc-devs at haskell.org
> http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.haskell.org/pipermail/ghc-devs/attachments/20250728/f1bb7196/attachment.html>