JSON transcripts off by two

[ct16k]

Hello,

Just wanted to say that the transcript in the info.0.json is from two comics ago (i.e. 1639/info.0.json describes #1637).

[myf]

Confirming for webpages as well: Looking at them with userstyle revealing transcripts [1], last comic with proper transcript was Hoverboard #1608, next two have their empty but in fact they appears shifted as described by OP. I.e. transcript for Food Combinations #1609 appears below Baking Soda and Vinegar #1611.

[1] userstyles.org/styles/84466/xkcd-show-transcript

[Ajedi32]

Confirmed. This problem is totally messing with the XKCD transcriber bot on Reddit.

[Hummin]

Confirmed again, if it helps. The bug seems to skip some transcripts, haven't worked out the algorithm yet.

[T-Rex]

It's also really annoying for xkcd apps like Easy xkcd. It has a search function that also checks the transcripts, but now gives wrong search results for the newer comics.

[myf]

Offset increased to three: e.g. transcript for 1674 is displayed under 1677.
I hope it is not the second kind of bug from 1700.

[OneTimeTranscriptCorrector]

Just to clarify if people read this later:

Code: Select all

if 1662 >= comicNumber AND comicNumber >= 1608 {
    transcriptComic = comicNumber + 2
} else if comicNumber >= 1663 {
    transcriptComic = comicNumber + 3
} else {
    transcriptComic = comicNumber
}


transcriptComic will be the comic with the correct transcript. Also note that comics after 1678 (inclusive) do not have transcripts (yet) and transcriptComic could be a comic that doesn't exist.

This means that (until this is fixed) comics 1608 (Hoverboard) and 1663 (Garden) do not have transcripts and are messing up all the other transcripts. These are the two interactive comics that are wierd, which is probably why it is broken.