On 14 Aug 2011, at 21:34, Christiaan Hofman wrote:
>
> On Aug 14, 2011, at 19:19, Markus Guhe wrote:
>
>> Dear BibDesk developers,
>>
>> First of all, apologies for sending this directly to the developers' mailing
>> list, but it seemed the least inappropriate thing to do ...
>>
>> I have a question about the performance of BibDesk AppleScripts in Lion. The
>> script below worked like a charm in OS X versions up to 10.6, but
>> performance got noticeably worse in 10.7. Admittedly, the script is a bit
>> twisted -- it looks for the record that contains a link to a particular file
>> path and returns the values of some fields of that record --, but
>> performance has never been an issue. I'm just wondering whether perhaps you
>> came across anything similar?
>>
>> If you like some context: I am using this script in my SlipBox application
>> to retrieve information from a BibTeX record when a user wants to import
>> annotations of a PDF that he/she made with Skim.
>>
>> Thanks,
>> Markus
>>
>> tell application "BibDesk"
>> set theDoc to document 1
>> tell theDoc
>> set pubList to publications
>> set thePath to "/Users/blah/Desktop/ABC.pdf"
>> repeat with thePub in pubList
>> tell thePub
>> set theFiles to get linked files
>> repeat with theFile in theFiles
>> set theURL to URL of linked file 1
>> if theURL is thePath then
>> set theCiteKey to the cite key
>> of thePub
>> set thePages to value of field
>> "Pages"
>> set theKeywords to value of
>> field "Keywords"
>> return {theCiteKey, thePages,
>> theKeywords}
>> end if
>> end repeat
>> end tell
>> end repeat
>> return {pubList}
>> end tell
>> end tell
>>
>> --
>> Markus Guhe
>> [email protected]
>> http://tabi-software.com/
>>
>
> I can't say anything about performance on Lion, as I don't have Lion myself.
>
> I also see nothing in the script that looks like it could be a bottleneck.
>
> But of course AppleScript support is 100% unpredictable and 100% unreliable,
> it never works exactly as it should work and never works or doesn't work the
> same way across OS versions.
>
> Apart from this, I am not sure what this script is supposed to do, but I am
> sure it does not do what you want it to do, because the main part does
> exactly nothing. Also, what it returns is completely inconsistent, when it
> finds nothing it returns a list of a list of pubs, otherwise it returns a
> list of some field values. Huh? The main point though is that the "if" block
> always will NOT be executed, because a URL is never equal to a path. Perhaps
> you want to check something like 'if (POSIX path of linked files contains
> thePath)'?
>
> Christiaan
Thanks for the quick reply!
Well, the script actually worked (except for the strange return value at the
end -- debugging leftover, sorry), meaning for some reason the if block has
been executed. Anyway, I changed it to use POSIX paths, which is much better,
of course, and it actually seems to run faster ...
Many thanks again and thanks for the great application!
Markus
------------------------------------------------------------------------------
uberSVN's rich system and user administration capabilities and model
configuration take the hassle out of deploying and managing Subversion and
the tools developers use with it. Learn more about uberSVN and get a free
download at: http://p.sf.net/sfu/wandisco-dev2dev
_______________________________________________
Bibdesk-develop mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/bibdesk-develop
