Enhancement: make it very easy to patch or use a non-patch method for the chkentry semantics tables

[xexyl]

Is there an existing issue for this?

• I have searched for existing issues and did not find anything like this

Describe the enhancement

A quite problematic issue with the semantic tables of chkentry is that if some new field has to be added or changed we have to jump through hoops to get chkentry(1) to not complain.

There have been numerous ideas we have floated but the most recent one (that both @lcn2 and I have thought of individually) is (as he wrote):

Instead a more descriptive file / file format should be used to indicate what needs to be changed. A reference JSON file can produce a table, but then something needs to modify that table to indicate that something might occur with a wider count range, for example.

Relevant images, screenshots or other files

No response

Relevant links

No response

Anything else?

Not yet.

[xexyl]

Although this is technically more of an enhancement I'm almost tempted to call it a 'bug': or at least a most annoying mis-feature. But of course it works well as long (as we don't have to modify anything) and something had to be done.

Of course working on this is post IOCCC28 but now we can at least discuss it (or not forget it as easily).

I do wonder if the jsemtblgen related code in jparse will have to change (I hope not) but I guess we'll see about that later on.

[lcn2]

We believe we have addressed all of the current questions that still need answering at this time. If we've missed something or something else needs to be clarified, please ask again.

[lcn2]

Important checkpoint

Please see GH-issuecomment-2762004417 for a recent comment on existing issues, priorities, and on a deadline for going into pending state for IOCCC29.

[xexyl]

Okay this issue is a hairy one ... it requires A LOT of thought and care and it might be better to have it in jparse as the semantics tables stuff is from jparse.

I am not closing this as a reminder but I am not sure how critical this is. I would say that this is not critical before IOCCC29. It is a nice to have but we both know how to do it otherwise (though I have to think about it every time it's still doable - although having to patch is still annoying and not very user friendly).

[lcn2]

TODO: When this issue is resolved, add a tables for the semantic check of .entry.json and then update chkentry.c to pass a non-NULL struct json_sem pointer to the chk_json() function.

[xexyl]

TODO: When this issue is resolved, add a tables for the semantic check of .entry.json and then update chkentry.c to pass a non-NULL struct json_sem pointer to the chk_json() function.

I believe that you were going to take care of this one as you have some ideas on it?

Earlier today I replied to your replies to my questions in the other issue.

With your answers to that I should be able to finish it. Although perhaps not tomorrow. I am not entirely sure how this coming week will go but most of the stressful stuff is about over.

Sleep is awful but I tend to be able to focus enough anyway.

Leaving now. I doubt I will do anything else here today.