Review TYPO3_4-5 Merged

4230[BUGFIX] Maintain compatibility with changed headline rendering

Helmut Hummel CR +2 V +2 Normal +96 −1 updated 06/03/13, 20:49

Commit message

If the fontTag property is set and the dataWrap property is set to the
default value, replace the dataWrap with the fontTag property value and
disable insertData on this level (if set).

This is to retain compatibility with versions before 4.5.4 while
compatibility with modified templates (before and after 4.5.4) is still
provided.

Review

  • Code review CR +2
    • Oliver Hader +2
    • Ernesto Baschny +1
    • Michael Stucki +1
  • Verified V +2
    • Oliver Hader +2
    • Michael Stucki +2

Issues 2

Files 2

Review log 29

Show 4 earlier messages
  1. Michael Stucki commented on patch set 1 08/12/11, 16:30

    Fails; I would prefer that you didn't submit this Hey Helmut, thanks for coming up with an alternative proposal. However, what this will basically do is to bring back the issue that users can read arbitrary data. So in any case, it would need further investigation. I have given some thoughts on the issue and will publish a new proposal based on my https://review.typo3.org/#change,4225 - michael
  2. Michael Stucki commented on patch set 1 08/12/11, 16:30

    Well, just like before the issue was fixed? :)
  3. Michael Stucki commented on patch set 1 08/12/11, 16:30

    No score; No score Ah, forgot my last note. Had another look at it, and of course the issue is not opened again. Will test your patch, and report back then...
  4. Michael Stucki commented on patch set 2 08/12/11, 16:30

    Patch 2 is almost the same approach but without adding any new TypoScript values. The (known) default values of dataWrap are matched against in tslib_content in order to detect if fontTag needs to be copied onto dataWrap or not.
  1. Michael Stucki commented on patch set 2 08/12/11, 16:30

    Patch 3 adds Olivers suggestions. Additionally I decided to remove insertData in such cases because it would definitely reintroduce the security issue which was already fixed. In theory this may again break a site if someone really intended to let users write getData values in the headline. However, unlike before the layout is no longer affected as only content would change...
  2. Michael Stucki uploaded patch set 2 08/12/11, 16:30

  3. Michael Stucki uploaded patch set 3 08/12/11, 16:30

  4. Michael Stucki uploaded patch set 4 08/12/11, 16:30

  5. Francois Suter commented on patch set 1 08/12/11, 16:30

    I would prefer that you didn't submit this I agree with Steffen: the change is out there, can't we just leave it and not introduce additional, extremely specific code? I think the solution is to comunicate even more, remind people of the change again when 4.5.5 is published, if possible across all mailing lists (with a pointer to the buzz article).
  6. Stefan Neufeind commented on patch set 2 08/12/11, 16:30

    Looks good to me, but someone else must approve Solution looks okay to me. Thanks for following up on this topic with so much energy.
  7. Oliver Hader commented on patch set 2 08/12/11, 16:30

    (2 inline comments)

    class.tslib_content.php:1998

    please add an additional isset($conf['dataWrap'])

    class.tslib_content.php:2004

    since this is magic, can we add a logging here? e.g. deprecation log or logging to $GLOBALS['TT']
  8. Oliver Hader commented on patch set 2 08/12/11, 16:30

    Looks good to me, but someone else must approve Seems to be working
  9. Michael Stucki commented on patch set 5 V 0 08/12/11, 16:30

    by team member; Looks good to me, but someone else must approve
  10. Ernesto Baschny commented on patch set 5 08/12/11, 16:30

    Looks good to me, but someone else must approve
  11. Oliver Hader uploaded patch set 5 08/12/11, 16:30

  12. Oliver Hader commented on patch set 5 08/12/11, 16:30

    Verified; Looks good to me, but someone else must approve I've added two test cases for this special configuration.
  13. Oliver Hader commented on patch set 5 V 0 08/12/11, 16:30

    by team member; Looks good to me, approved
  14. Xavier Perseguers commented on patch set 1 08/12/11, 16:30

    As one has to read it slowly and really think about it to guess what it does, it really lacks some proper documentation how it works.
  15. Helmut Hummel commented on patch set 1 08/12/11, 16:30

    @Xavier: what exactly do you miss? @Steffen: Please post a TS setup that breaks with this patch
  16. Helmut Hummel commented on patch set 1 08/12/11, 16:30

    @Steffen: ok, now I got it. Now I'm out and back to my first impression that this cannot be fixed in a reasonable way, but we have to accept the breaking change (which is out now anyway).
  17. Helmut Hummel commented on patch set 1 08/12/11, 16:30

    Hey Michael, how could they read arbitrary data with that patch?
  18. Steffen Ritter commented on patch set 1 08/12/11, 16:30

    > Now I'm out and back to my first impression that this > cannot be fixed in a reasonable way, but we have to accept > the breaking change (which is out now anyway). full ack, let it as is - people affected will already have fixed it...
  19. Steffen Ritter commented on patch set 1 08/12/11, 16:30

    I would prefer that you didn't submit this Sorry Helmut, I really do not like such "Magic Code". Furthermore this does not ensure that this will fix it? People who adapted the header rendering might not using fontTag for it... So what happens if they already did fonttag> and used another wrapper? Problem resides... Therefore I really opt NOT to add such magic. Just place "release more buzz" about it. Probably we should use template analyzer to detect if someone fiddled with lib.stheader and just report a warning...
  20. Mr. Jenkins commented on patch set 1 08/12/11, 16:30

  21. Mr. Jenkins commented on patch set 2 08/12/11, 16:30

  22. Mr. Jenkins commented on patch set 3 08/12/11, 16:30

  23. Mr. Jenkins commented on patch set 4 08/12/11, 16:30

  24. Mr. Jenkins commented on patch set 5 08/12/11, 16:30

  25. Oliver Hader merged patch set 5 08/12/11, 16:30