-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Document 001: lists in paragraphs (in XML) #341
Comments
(This causes a breakage in Firelight build, which does not allow non-flow content in paragraphs, and while it handles embedded notes and footnotes it does not handle lists [yet?]. Related: metanorma/firelight#46) |
hello @strogonoff thank you for pointing this out. i am not familiar with the paragraph id. if you can tell me where it is (which and clause and perhaps some text) i will check it. it seems like its an anomaly so it can probably be corrected in the document content. thank you. |
hello @strogonoff and @opoudjis i looked in the this is an autogenerated UML table. the source .adoc code for this is coming out of the Enterprise Architect (EA) after being input in the "Notes property" diagram. an .xmi file is out from EA and lutaml reads that and the contents are autogenerated into these tables. i dont have the current model. if you can describe what should be changed, i will ask the client to update the .adoc code in the model. i indicated to the client that complex adoc "structures" are not be supported. if this is a complex adoc "structure" which mn does not support the text in the EA model can be modified. please let us know. thank you. |
@ReesePlews I worked around this problem. This is primarily a bug for Nick, because according to him Metanorma should handle lists in other ways. However, I see that you merged some changes that also fixed the build. Maybe you removed that use of lists? I’m not sure. Either way, checking the build now. |
thank you @strogonoff ; there could be an issue in the model, which is where the .adoc is stored for these type of "autogenerated tables" ; if changing that .adoc code is less of a problem for @opoudjis we should consider that from the standpoint of a mn code fix and testing compared to some simple modification of the content .adoc code. let me know what you both feel is better in this case(s). thank you. |
I believe we have no list-related issue now. |
Not sure if @opoudjis wanted to keep it open for his purposes. Either way, Firelight doesn’t break on it. |
hello @strogonoff i checked the output in a local build of doc01 (sources/001-v5) but the list from the PDF is still an ordered list of a, b, c and then an un-ordered list of bulllets, identical to the first image. i agree we should keep this open to see if this fix will be in the forthcoming metanorma release. thank you. |
As I have REPEATEDLY said in the past, I cannot grep for screencaptures, and I am do not have the time to work through the convoluted processes you have in place for generating Asciidoc. I expect you to be more helpful than this. Based on the underspecified indication above, I gather the source in the XMI document is:
THIS should have been pasted into the ticket. |
Translation:
|
The XML that is generating is:
which does not have the nesting claimed for paragraphs. However, @ReesePlews,
that is EXACTLY what the Asciidoc specifies:
So Reese, this is not a bug, this is what was written in the XMI file; and Anton, you or whoever is maintaining metanorma-plugin-lutaml need to provide me with the Asciidoc that is generating whatever XML you are seeing, and for that matter what XML you are seeing. Without adequate data to action a bug report, I will not be actioning bug reports. |
@strogonoff No idea what to do with that. I need XML I can access. |
@opoudjis I believe @strogonoff meant this "mn" artifact, not the "artifact page" which contains a failed Firelight build of 001: This is where the XML is. |
I have confirmed that Anton is seeing the XML with ordered lists nested inside of paragraphs, and I cannot replicate that based on the source Asciidoc. I am generating the document locally, dumping the preprocessor output to disk, so I can see what exactly Asciidoc is being processed. My inclination is to do that universally, this keeps coming up as a debugging requirement as we keep using preprocessors like lutaml. |
There is the mn artifact in the build summary. I worked around this issue, so it does not break Firelight. |
The processing of those XMI tables happens in Lutaml Table processing, which is block processing, so I cannot get to it in preprocessing, as it hasn't happened there yet: I need to grab it at the start of Metanorma processing proper:
|
No, that is not catching the results of Lutaml Table processing. This is EXCEEDINGLY difficult to debug, and I needed this like a hole in the head when I'm in the middle of refactoring. It doesn't matter that Anton has a workaround, this is clearly contaminated Asciidoc, and there will be wrath doled out if I find out that it is not my fault. One thing that is occurring to me is the monstrously dumb way Asciidoc has been shoved into Enterprise Architect as an XML attribute. An XML attribute, of all things. Its linebreaks are being done as All it would take is for the code processing that Enterprise Architect XMI to mangle the linebreak, for the Asciidoc to turn out dodgy. |
I STILL cannot replicate this locally!!!
is giving me properly nested XML. And look at the HTML: It has spaces between blocks. What Reese is seeing does not.
This looks to me like something that @kwkwan needs to investigate in lutaml_klass_table under metanorma-plugin-lutaml . Asciidoc with non-native linebreaks is incorrect Asciidoc, and that needs to be rectified at the source, it is too late by the time Metanorma gets to it. |
The difference is this: Correct: (generates valid XML)
Incorrect: (generates invalid XML)
The weird part is the “incorrect” example used to always merge the lines into the same paragraph, ie there is no list detected and therefore no invalid XML. I don’t understand why the behavior has changed. |
We should have a place on our website about known issues and workarounds. |
Is this worthy of some sort of metanorma/coradoc linting tools, to detect "doc-smell" like this? (existing implementation: https://github.com/docToolchain/asciidoc-linter) |
Hi @ReesePlews and @strogonoff I have updated the liquid template of the class table to replace |
thank you @kwkwan i will check later today. if there are any other adjustments to make in the asciidoc code please suggested them and i will discuss with the client. @opoudjis regarding the XML fragments. i am sorry i just dont understand enough about what XML to give you to track down the issue. i have been trying to add links to adoc files and other clarifications (based on earlier advice from you) but this was the first time dealing with checking the actual XML to find this issue. i am sorry it was not enough. |
In document 001’s XML (can be found in this artifact), the paragraph with ID of
_c0ef23ff-7174-e24b-2402-0b7829592fc7
contains a list within. It seems like it may be a bug according to @opoudjis.The text was updated successfully, but these errors were encountered: