Forum for BQM 450

Regarding "force regeneration of diagram" [#1703]

Back to forum     Back to page

New reply

 
reply
Unkx80: Regarding "force regeneration of diagram" (2009-01-08 20:36) [#5616]

Earlier Andy made a remark on the main page: why doesn't $$Bm11 make black's first move be #11 in this diagram? (this refers to the "continuation" diagram). Then I forced the regeneration of this diagram and viola, the first move became #11. Subsequently I deleted that remark. I don't know why #11 failed to show in the first place.

I forced regeneration of the diagram by making an arbitrary change to it and saving it, and then editing it again to undo the change and saving it again. There is an effect on the diagram after each save, but it did not appear in the diffs as consecutive edits are treated by the wiki as one single edit.

For those interested, I know that the string of characters representing a diagram is mapped to a MD5 hash. So internally, the wiki manages diagrams using these MD5 hashes as some kind of identifiers. More specifically, the wiki generates a diagram if no diagram exists with its corresponding MD5 hash, and deletes a diagram when it determines that no page refers to it.

X
fractic: Re: Regarding "force generation of diagram" (2009-01-08 20:33) [#5617]

This has happened before. See [ext] http://senseis.xmp.net/?topic=1426.

BTW. can you link directly to a forum topic rather then using an external link?

Unkx80: Re: Regarding "force generation of diagram" (2009-01-08 20:39) [#5618]

Yes. topic:1426 or t:1426. More at text formatting rules under "References (Links)".

DaveSigaty: Re: Regarding "force regeneration of diagram" (2009-01-09 00:51) [#5619]

Does this mean that the top line of the diagram is not included in the hash? If it were, the edit that added "m11" to the top line should have changed the hash and caused a regeneration of the diagram, right? Is this a bug?

fractic: Re: Regarding "force regeneration of diagram" (2009-01-09 01:01) [#5620]

In the earlier topic Arno said that the first line was taken into acount so it should work. Clearly something is wrong though.

ArnoHollosi: Re: Regarding "force regeneration of diagram" (2009-01-09 07:59) [#5621]

Here are two diagrams that are exactly the same but for the move number. If you have a look at the name of the SGF or PNG you see that they are different - i.e. the move number is part of the hash. Unkx80 described the mechanism of diagram creation/deletion rather well. I honestly have no idea what could lead to the described behaviour or if it really is a bug. It is not reproducible.

If you want to give me chance to debug it, then the next time it happens do not force the regeneration, but leave the erroneous diagram there.

[Diagram]
19x19 diagram  
[Diagram]
19x19 diagram  
 
Back to forum     Back to page

New reply


[Welcome to Sensei's Library!]
RecentChanges
StartingPoints
About
RandomPage
Search position
Page history
Latest page diff
Partner sites:
Go Teaching Ladder
Goproblems.com
Login / Prefs
Tools
Sensei's Library