List numbers wrong for skeleton output

[denismerigoux]

When running catleg skeleton LEGITEXT000006069577 LEGISCTA000006179571, for "Article 13" the first 3 enumerated items are 1. instead of 1., 2. and 3.

[rprimet]

good catch!

[denismerigoux]

So for some reason there's still a problem in the French income tax lint with these list numbers : https://gitlab.adullact.net/dgfip/ir-catala/-/jobs/191411

[rprimet]

@denismerigoux was the skeleton generated with the latest catleg version?

Trying catleg article LEGIARTI000044983201 seems to yield the correct numbering?

[rprimet]

(if the skeleton was generated using a previous version, then the diff might correctly exhibit the former bug?)

[denismerigoux]

Trying catleg article LEGIARTI000044983201 seems to yield the correct numbering?

Locally on my machine yes... Maybe it's a problem with fetching the latest catleg version here ? https://gitlab.inria.fr/verifisc/docker-catala/-/blob/master/Dockerfile?ref_type=heads#L20

[denismerigoux]

@AltGr are you sure the Catala CI image gets an updated catleg with https://gitlab.inria.fr/verifisc/docker-catala/-/blob/master/Dockerfile?ref_type=heads#L20 ?

[rprimet]

@denismerigoux @AltGr it's possibly not a CI problem, more like which version was used to (hand-)generate the skeleton that was then committed during a pair-programming session (the CI seems to be doing its job fine)?

(i.e. the problem was in a former version of catleg skeleton which is used during programming, and not in catleg diff which the CI runs)

[denismerigoux]

No, the version in the source code against which catleg lints is fine : https://gitlab.adullact.net/dgfip/ir-catala/-/blob/main/sources/cgi_revenus.catala_fr?ref_type=heads#L65-112 (the diff from the lint job is here https://gitlab.adullact.net/dgfip/ir-catala/-/jobs/193158#L25).

[rprimet]

OK, thanks, I can reproduce the issue now