Fail a license compare if there is text following a matching license

[goneall]

When using the LicenseCompareHelper.isTextStandardLicense and there is text in the compare license text following a valid license text match, fail the license compare.

This issue was originally raised in spdx/license-list-XML#1583

Signed-off-by: Gary O'Neall [email protected]

[goneall]

Note that this fix is not yet complete. It is failing one of the regression tests where a license template ends with optional text.

[goneall]

The following license fail to match against the text after this fix and should be investigated and resolved:

• Python-2.0.1 - Pull request submitted - Removing 0BSD license from Python 2.0.1 test file license-list-XML#1615 (see comment for context 0BSD license in Python-2.0.1.txt license-list-XML#1583 (comment))
• HPND - Pull request submitted: Fix Alt regular expression in HPND license-list-XML#1587
• APSL-2.0 - This template and text looks correct, this could be an issue with the license matcher Pull request created: Fix APSL-2.0 exhibit text to match test code license-list-XML#1593
• APL-1.0 - This template and text looks correct, this could be an issue with the license matcher Pull request created: Fix matching problem for APL-1.0.xml license-list-XML#1594
• CPAL-1.0 - This template and text looks correct, this could be an issue with the license matcher Pull request created: Fix matching error to original license for CPAL-1.0 license-list-XML#1595
• Fair - Extra line at the end of the test file - Pull request submitted Remove extraneous line at the end of the fair text file license-list-XML#1589
• GFDL-1.1-invariants-or-later - Note that the GFDL-1.1-invariants-only matches. The templates differ in that the last optional in the or-later version has a bunch of extra matching alt text. There is most likely an error on one of these regexes. Pull request submitted Fix bad match pattern GFDL-1.1-invariants-or-later license-list-XML#1621
• MPL-2.0-no-copyleft-exception - This template and text looks correct, this could be an issue with the license matcher Pull request created: Remove extraneous > from MPL-2.0-no-copyleft-exception license-list-XML#1596
• OSET-PL-2.1 - This template and text looks correct, this could be an issue with the license matcher Created PR: Fix OSET-PL-2.1 matching issue license-list-XML#1597
• SGI-B-1.0 - This template and text looks correct, this could be an issue with the license matcher Based on other non-matching licenses, this is likely due to an error in the LicenseXML alt text in the last optional field (starting with Exhibit A). It would take quite some time to debug this, suggest we just remove the complex matching and just include the original text. Fixed with commit a2450dc
• SGI-B-1.1 - This template and text looks correct, this could be an issue with the license matcher Based on other non-matching licenses, this is likely due to an error in the LicenseXML alt text in the last optional field (starting with Exhibit A). It would take quite some time to debug this, suggest we just remove the complex matching and just include the original text. Fixed with commit a2450dc
• Spencer-86 - Extra line in the test file - pull request submitted: Remove extra line in Spencer-86 test file license-list-XML#1619
• SPL-1.0 - This template and text looks correct, this could be an issue with the license matcher Created pull request: Match test text or SPL-1.0 license-list-XML#1598
• SSPL-1.0 - Submitted issue: Extra line in SSPL-1.0 test file license-list-XML#1591
• SugarCRM-1.1.3 - This template and text looks correct, this could be an issue with the license matcher Pull request submitted: Remove extraneous comma from SugarCrM-1.1.3 license-list-XML#1599
• gnu-javamail-exception - Created pull request: Remove extraneous quote from gnu-javamail-exception test file license-list-XML#1592

[goneall]

Note that almost all of the test failures due to the license matcher has an Optional at the very end of the XML file (most are optional Exhibits). Also note that the error indicates the missmatch is the first word after the start of the optional text.

[goneall]

@swinslow - looks like all of the errors in the LicenseListXML comparison are due to errors in the XML markups.

I fixed all exception GFDL-1.1-invariants-or-later, SGI-B-1.0, and SGI-B-1.1. For these, I think we should just remove the extra matching stuff in the last optional section. It would probably take 3-4 hours to get these more complex markups to work. Let me know what you think.

[swinslow]

Thanks @goneall! This is awesome. Thanks for submitting all the fixes in the license-list-XML repo.

I will try to take a closer look at this soon. I'm about to be offline for a while, but will aim to review the PRs and the three remaining ones that you mentioned when I'm able to do so.