BlockStatement not migrated

[Turbo87]

I don't know why, but the following snippet was not converted correctly. The code lived inside another block that was correctly converted.

          {{#rl-dropdown-toggle tagName="a" class="dropdown"}}
            {{svg-jar "sort"}}
            {{ currentSortBy }}
            <span class='arrow'></span>
          {{/rl-dropdown-toggle}}

[tylerturdenpants]

Wooo a new bug to squash. I’ll try to take a look at it this weekend.

[Turbo87]

@tylerturdenpants this was on https://github.com/rust-lang/crates.io btw

[tylerturdenpants]

@Turbo87 I spoke too soon. Deleting previous comment.

[tylerturdenpants]

Ok, so I was right, because crates.io runs [email protected], there are functions missing in the meta that fail to be captured during enumeration. I will add a guard in telemetry helpers for these functions

[Turbo87]

hmm, good point. I'm not sure in which scenario I tried to run the codemod, because I do have a branch that updates the project to 3.15 😅

[Turbo87]

hold on, this is a block statement, can that be anything other than a component invocation? (except for if/each)

[tylerturdenpants]

I can give you the DEBUG info in a few hours to show you that although the module for the component is there, evaluating it causes an error, and therefore never makes it to the component types array in the telemetry

[tylerturdenpants]

Plus the telemetry code doesn't certify a particular block mustache statement as valid component invocation even though it is. I'd have to enhance the codemod to apply these heuristics. A PR such as this would be awesome from a skilled programmer such as yourself... 😁

[Turbo87]

yeah, I think we only should rely on the telemetry data in ambiguous cases :)

[rwjblue]

Agree.

[tylerturdenpants]

@rwjblue I agree too... but I'm trying to find the time! 😄 Busy holiday stuff, work, and family.