Stream: IG creation
Topic: New errors on Change history generated content
Jean Duteau (Jul 28 2020 at 19:01):
@AbdulMalik Shakir had the same errors as I am reporting, but he reported them in the "provenance history not working" thread and I thought I'd break it out into a separate thread. The Change History page that is being generated is rife with broken links:
ImplementationGuide-hl7.fhir.us.davinci-pas.change.history.html#/html/body/div/div/div/div/div/ul/li/a at Line 186, column 6 error The link 'ImplementationGuide-hl7.fhir.us.davinci-pas.html' for "Narrative Content" cannot be resolved
Eric Haas (Jul 28 2020 at 19:43):
@Jose Costa Teixeira ?
"ImplementationGuide": {
"template-base": "",
"template-format": ""
},
This says there will not IG resource format or base pages created but the Hx page will still have tabs pointing to them it is generated. its all or nothing.
need to add in
"ImplementationGuide": {
"template-base": "",
"template-format": "",
"template-change-history":""
},
Jose Costa Teixeira (Jul 28 2020 at 19:45):
will try (even if the history pages should not be generated because the resources don't have a history
Jean Duteau (Jul 28 2020 at 19:47):
i'm 99% confident that Eric is right
Eric Haas (Jul 28 2020 at 19:48):
I'm 99% sure is a template thing that cropped up before and I fixed and Jose unfixed
Eric Haas (Jul 28 2020 at 19:49):
any page that is generated will have tabs that point to other pages
Eric Haas (Jul 28 2020 at 19:51):
although Jose is correct that the history page is being generated mysteriously
Eric Haas (Jul 28 2020 at 19:59):
Jose: What is happening is we have this parallel set of pages defined in the config.json
extension-blah.xml.html - these pages are being pointed by the Profile history page tabs but not being generated for profiles
extension-blah.profile.xml.html - these are being generated and should be the the ones pointed to by the hx page.
Eric Haas (Jul 28 2020 at 20:04):
so the history pages being created are using the StructureDefinition-extension-blah.change.history.html ie not the profile specific pages...
Eric Haas (Jul 28 2020 at 20:04):
and Jean is correct once again....
Jean Duteau (Jul 28 2020 at 20:05):
dang, I just went and edited my post to say that you were right!
Jose Costa Teixeira (Jul 28 2020 at 20:18):
@Eric Haas you mean this?
"template-profile-xml": "template/layouts/layout-profile-format.html",
"template-profile-json": "template/layouts/layout-profile-format.html",
"template-profile-ttl": "template/layouts/layout-profile-format.html",
"profile-xml": "{{[type]}}-{{[id]}}.profile.xml.html",
"profile-json": "{{[type]}}-{{[id]}}.profile.json.html",
"profile-ttl": "{{[type]}}-{{[id]}}.profile.ttl.html",
Jose Costa Teixeira (Jul 28 2020 at 20:20):
is this the unfix you mean? https://github.com/HL7/ig-template-base/commit/a112a46edf1c29777a668d53dbe45e7a7ffe43c7
Jose Costa Teixeira (Jul 28 2020 at 20:29):
Eric Haas said:
Jose: What is happening is we have this parallel set of pages defined in the config.json
extension-blah.xml.html - these pages are being pointed by the Profile history page tabs but not being generated for profiles
extension-blah.profile.xml.html - these are being generated and should be the the ones pointed to by the hx page.
I don't remember if the previous issue was related. In this case, these history pages should not be generated.
Jose Costa Teixeira (Jul 28 2020 at 20:36):
I'm 99% confident this is very weird.
Jose Costa Teixeira (Jul 28 2020 at 21:14):
i can not fix both issues. the first one seems to be fixed with @Eric Haas 's fix
Jose Costa Teixeira (Jul 28 2020 at 21:14):
@Lloyd McKenzie PR created
Jose Costa Teixeira (Jul 28 2020 at 21:16):
Sorry Eric and others, indeed apparently I had removed that line while fixing the other issues
Jose Costa Teixeira (Jul 28 2020 at 21:16):
this is what happens when I "fix" issues before drinking
Jose Costa Teixeira (Jul 28 2020 at 21:19):
about the second issue (structuredefinitions getting a history page) - I presume @David deRoode you are using sushi?
David deRoode (Jul 28 2020 at 21:23):
No
Jose Costa Teixeira (Jul 28 2020 at 21:29):
nevermind. I think I found it.
Jose Costa Teixeira (Jul 28 2020 at 21:50):
For checking: @Lloyd McKenzie @Eric Haas etc
I was under the impression that the "Any" applies to "any other type not specified", but apparently "All" sets attributes to "All types".
When I changed the history page names, I also added a default template for all other types (assuming that this would not impact the types that have their section there). So I just added the settings to clear that.
PR submitted
Eric Haas (Jul 28 2020 at 22:04):
I was under the impression that the "Any" applies to "any other type not specified", but apparently "All" sets attributes to "All types".
Where is the All attribute?
Eric Haas (Jul 28 2020 at 22:18):
I think we have to define the
"template-change-history":
for "StructureDefinition:extension":
and "StructureDefinition":
as well so that there are both remplate-profile-history and template-change-history pages present.
(or remove this duplication replaced with some extra liquid in single template)
Jose Costa Teixeira (Jul 28 2020 at 23:13):
Eric Haas said:
I was under the impression that the "Any" applies to "any other type not specified", but apparently "All" sets attributes to "All types".
Where is the All attribute?
I mean "Any" (tried to fix above but I can't)
Jose Costa Teixeira (Jul 28 2020 at 23:15):
Eric Haas said:
(or remove this duplication replaced with some extra liquid in single template)
if you can do the liquid stuff that would work. The history pages are different and I think this is also related to the nav-tabs..
Jose Costa Teixeira (Jul 28 2020 at 23:16):
Eric Haas said:
I think we have to define the
"template-change-history":
for "StructureDefinition:extension":
and "StructureDefinition":
as well so that there are both remplate-profile-history and template-change-history pages present.
I don't see why. The 2 layouts are just named differently because they have slightly different content and navigation options. The profiles use one layout, the rest uses the other
Jose Costa Teixeira (Jul 28 2020 at 23:17):
@David deRoode is the issue gone?
Last updated: Apr 12 2022 at 19:14 UTC
