Successful return code on encoding error

Quick method to test:
sudo apt-get install html2text
curl -s https://launchpadlibrarian.net/250515419/test.html | html2text; echo $?

The simplest fix is to replace the "continue" in the recode error handlers with "exit(1)" (Patched html2text.C lines 595 and 555). (Using a different value for encoding errors might be useful)

This makes the error handling consistent with other cases, like failing to open a file. (See line 482)

A better solution would be to keep track of whether an error occurred in the loop and then change the final return value based on that. (Requiring changes to add a variable declaration, and on lines 482, 510, 555, 582 and 595)

(Line numbers are from the patched Wily version)

My attempt at an actual patch.

The other method, of actually changing all the error handling to save that there have been an error and returning the code at the end might still make sense for later.

The attachment "Patch to modify patches to return an error code on recoding errors" seems to be a patch. If it isn't, please remove the "patch" flag from the attachment, remove the "patch" tag, and if you are a member of the ~ubuntu-reviewers, unsubscribe the team.

[This is an automated message performed by a Launchpad user owned by ~brian-murray, for any issues please contact him.]

Note: This is probably also present in Debian, since it seems like the Ubuntu package is mostly unmodified.

Also affects 1.3.2a-18 on Ubuntu 16.04.

Status changed to 'Confirmed' because the bug affects multiple users.

I can confirm the same incorrect response on Ubuntu 14.04.5 LTS with all updates installed.

I logged this upstream at https://github.com/Alir3z4/html2text/issues/156 as well.

@Roland: Incorrect upstream, that is python-html2text / python3-html2text. This is from http://www.mbayer.de/html2text/ (And that won't have the issue, since it was introduced in Debian's UTF-8 patch) (Debian might also be affected though)

Gert, you're right, my bad.

I get the full picture now. Have you reported this to the Debian bugtracker?

Confirmed on Debian 8:

$ curl -s https://launchpadlibrarian.net/250515419/test.html | html2text; echo $?
Input recoding failed due to invalid input sequence. Unconverted part of text follows.
“Test��?)
</body>
</html>

0

I haven't reported it to Debian, I did not have a test system to confirm that it is present there as well (I'm quite sure it would be present, it is the same package source) (And https://www.debian.org/Bugs/Reporting for manual reporting seems quite scary...)