An simple example of of outputting a clpBNR interval variable with format/2: ?- {X>=1}, format('interval(~w)',X). interval(_2428{real(1,1.0Inf)}) X::real(1, 1.0Inf). If I run the same example on a local pengine using a test predicate local_query/1, the received output message doesn’t contain the a…

So all is well for formatted output of attributed variables if the write_attributes flag is set to portray in the query pengine, which also means it has to be declared a “safe” flag. Note this doesn’t appear to affect the top level answer format of attributed variables, presumably because it’s not r…

[image] ridgeworks: Note this doesn’t appear to affect the top level answer format of attributed variables, presumably because it’s not running in the query pengine(?). The SWISH toplevel printing goes through library(pengines_io), which eventually uses library(http/term_html), term//2 to ren…

[image] jan: The first questions is why can’t clpBNR support copy_term/3 , so it is in sync with the other constraint libraries? I believe it does in so far as it implements attribute_goals//1 which is required to support copy_term/3. So if this is the basis, for generating the answer output i…

[image] ridgeworks: I believe it does in so far as it implements attribute_goals//1 which is required to support copy_term/3. So if this is the basis, for generating the answer output in SWISH , I think all should be fine. :+1: [image] ridgeworks: I’m not sure how this would be supported …

[image] jan: Possibly you can either reuse this, add another hack for clpBNR or, better, try to generalize it. I hope you installed a local copy of SWISH to experiment with? So this is a complication I had hoped to avoid and will take some effort. [image] jan: Just submit a PR with the o…

[image] jan: The relevant stuff is mostly in lib/trace.pl of the swish sources. This essentially packs additional information such as the well founded semantics delay list in additional variables that it adds to the projection (bindings). From swish’s lib/trace.pl: %! pre_context(Name, Goal,…

[image] ridgeworks: write_attributes Added. [image] ridgeworks: Does the current clpBNR git structure meet the requirement, or am I missing the point. clpBNR is (I think( structured as a package, so that should be fine. There are roughly two ways to deal with packages. One is to i…

[image] jan: [image] ridgeworks: write_attributes Added. Thanks. I see it’s in SWIP 9.1.20 so I’ll do my final testing on thta. [image] jan: clpBNR is (I think( structured as a package, so that should be fine. The package actually contains two modules - clpBNR and clpBNR_utilit…

[image] ridgeworks: The fix is fairly simple and at some point I’ll try to generate a PR. :+1: [image] ridgeworks: If this is the solution, I’ll give a hard pass. I don’t know. There could be others. So far, the most problematic systems have been CHR and sCASP. Both because the…

[image] jan: clpBNR is (I think( structured as a package, so that should be fine. There are roughly two ways to deal with packages. One is to install them as user. The other is to make them submodules (provided they are available as a git module) of the main project and install them locally. Tha…

Using `format/2` with attributed variable on a pengine

Help!

ridgeworks December 7, 2023, 8:42pm 27

Got stuck so opened a new topic:

Topic		Replies	Views
Use of `attribute_goals//1` to compose answers Help!	41	822	December 19, 2023
Inspecting programs with attributed variables using clause/2 General	54	502	August 22, 2026
clpBNR flag management Discussion	22	844	October 14, 2022
Problem with pengine_self? Help!	14	651	June 3, 2019
Ann: SWI-Prolog 8.4.0 Releases ssu	2	1402	October 10, 2021

Using `format/2` with attributed variable on a pengine

Related topics