[jira] [Comment Edited] (CALCITE-4238) Create a default parser configuration, to reduce redundant information in sub-parsers

["Ruben Q L (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CALCITE-4238?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17204573#comment-17204573 ] 

Ruben Q L edited comment on CALCITE-4238 at 9/30/20, 9:30 AM:
--------------------------------------------------------------

[~julianhyde] sorry to disturb you. Should we be concerned about the AppVeyor failure that happened in the PR and has continued after the merge into master?
https://ci.appveyor.com/project/ApacheSoftwareFoundation/calcite/builds/35473782/job/sqi1w4g9a5ots03a

UPDATE: I have the impression the issue is caused by this line (which was modified by the current ticket):
https://github.com/apache/calcite/blob/539807bb0558890c2a92d0fc2af20c741a775fbc/buildSrc/subprojects/fmpp/src/main/kotlin/org/apache/calcite/buildtools/fmpp/FmppTask.kt#L66


was (Author: rubenql):
[~julianhyde] sorry to disturb you. Should we be concerned about the AppVeyor failure that happened in the PR and has continued after the merge into master?
https://ci.appveyor.com/project/ApacheSoftwareFoundation/calcite/builds/35473782/job/sqi1w4g9a5ots03a

> Create a default parser configuration, to reduce redundant information in sub-parsers
> -------------------------------------------------------------------------------------
>
>                 Key: CALCITE-4238
>                 URL: https://issues.apache.org/jira/browse/CALCITE-4238
>             Project: Calcite
>          Issue Type: Bug
>            Reporter: Julian Hyde
>            Assignee: Julian Hyde
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: 1.26.0
>
>          Time Spent: 0.5h
>  Remaining Estimate: 0h
>
> Create a default parser configuration, to reduce redundant information in sub-parsers. In particular, each sub-project parser would not need its own list of non-reserved keywords (currently the list is over 250 keywords, and the list is identical in each sub-parser).
> Since CALCITE-2405 each sub-parser has its own list of non-reserved keywords. This is good, because it allows the sub-parser to make keywords non-reserved (e.g. in DRILL-1065 Drill wanted {{EXEC}} [to be non-reserved|https://issues.apache.org/jira/browse/DRILL-1065?focusedCommentId=14903669&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-14903669]) but it still requires each parser to have a list of non-reserved keywords (see e.g. [the list in Babel parser|https://github.com/apache/calcite/blob/eb4011bb4b9d962c2f8e6767cbee8c23435ef939/babel/src/main/codegen/config.fmpp#L37]). And if we add a keyword to Calcite, we almost always have to modify the list in each sub-parser.
> This proposal would move the list of default non-reserved keywords to a shared config file. This would reduce the amount of code in sub-parsers, and remove the need to edit them when we add a keyword to the base parser.
> Also, we would not need to modify each sub-parser each time we add a markup variable (e.g. when we added {{binaryOperatorsTokens}} and {{extraBinaryExpressions}} in CALCITE-2843).



--
This message was sent by Atlassian Jira
(v8.3.4#803005)