On my plate
On my plate, in my browser tabs:
- /~distler/blog/files/MTStripControlChars.pl
- Musings: MTStripControlChars
- Sketchbook: m[iA]cro: On NoHTMLEntities and application/xhtml+xml
- Sam Ruby: Character Encoding and HTML Forms
- Survival guide to i18n
- Bug 228779 - Submitted characters not included in the iso-8859-1 charset for iso-8859-1 documents should be always encoded as numeric character references
- Bug 18643 - add support for _charset_ field in form submissions [form sub]
- Bug 81203 - smart quote cannot be submit in html form
- Windows Code Page 1252
- Web Forms 2.0
- MovableBLOG: The Real Problem With TypeKey
- i d l y . o r g :: A World Without TypeKey
- SemErgence: A Trust Web Might Help
- BBC NEWS | Anti-Barbie becomes
Russian icon - Scobleizer: Microsoft Geek Blogger (persuasive about persuasion)
- Whole Wheat Radio, Talkeetna - Alaska
All of them calling out for a full post, with plenty of thought, or testing, or both. Not a one that’s willing to let me get away with a quick link.
On my plate, in my inbox:
- My BlogThis Mozilla extension is stupidly broken, only passing through 255 characters of selected text.
- My PHP blo.gs blogroll script is probably stupidly broken with character encoding problems, and at any rate desperately needs to be rewritten.
- Revision Plugin wants to be blogged, needs to be hacked and patched if I’m right in thinking that it’s saving revisions while an entry is a draft.
- PreviousNextInCategory Plugin needs to be destupified, so that it defaults to only using the primary category, with an attribute to ask for the current weird any-category effect.
- OpenPGPComment 1.5 - I need to try updating, see if it will work without my current brutal hacks, and either way send what I’m doing off to Michael, who’s wondering whether I’m obnoxious about not answering email, or just really slow and stupid about it (answer: both).
- Beg Shannon to come back and play with us. Drive back there and beg in person, if that’s what it takes. I’m starting to get desperate.
I don’t suppose that spending ten minutes making a list of links actually got me any closer to getting any of those done?
Hrm. You’re right. I never bothered to test how the Revision Plugin works when dealing with drafts. Hrm. Hrm.
Should be easy, right? I just figure out how to tell if it’s a draft, and skip the checkin in that case. Hrmmmmm.
Should be easy? When has anything ever been easy? I went to look up the status syntax in the place where I know it best, in CMS.pm where it’s deciding whether or not to send a ping (first place I ever hacked MT), and saw:
if ($obj->status == MT::Entry::RELEASE() ||$status_old eq MT::Entry::RELEASE())
(where it’s deciding whether to rebuild the entry after saving. If you happen to need to use it, draft is
MT::Entry::HOLD()). But what about $status_old? That’s where it rebuilds when you change from RELEASE() to HOLD() - I start an entry as a draft, take out half the swearing and ad hominems, publish it (where you get your first copy), decide it’s too nasty, take it back to draft, take out another half, then publish it again. Do you keep the first public version, and diff with the second public version, skipping the drafts? Or does drafting a published post mean ”revision plugin, please remove all your history of this post”?Well, the spirit of the revision plugin is to track what has been publicly available in the past. So what I’d like it to do is check in when it’s published. So, publish first - checks in. Take to draft, nothing happens. Save draft, nothing happens (repeat many times as draft.) Then edit the draft one more time, and publish. It checks it in again as revision 2. A diff would be from the current version to the last public version, as those are the only two revisions in the rcs file.
To want to hide evidence of what was publicly available, then it’s the same solution as before - log in to the server and remove the rcs file.
Yes, if an entry were taken from publish to draft, its rcs file would still exist. But there would be no links around to click on to load it in. People could surely construct the links, but there’s enough of an obfuscation barrier there that I don’t think that’s a big concern, considering the user can always go remove the rcs file if they really want to wipe out evidence of publicly available data ever existing.
Seems to me like you would want to keep draft revisions as well as public revisions, the one caveat being that you might not want to display the draft revisions to the world at large. Simple solution, tag the document at publish time, and don’t publicly display anything before that tag.
Just noticed its built (quite logically) on RCS not CVS. I don’t know if RCS has tags.