[Date Prev][Date Next]
[Thread Prev][Thread Next]
[Date Index]
[Thread Index]
[New search]
To: <tmi@xxxxxxxxxxxx>, "FrameMaker user forum" <framers@xxxxxxxxxxxxxx>, <framers@xxxxxxxxx>
Subject: Search index not available on Solaris
From: "Peter Ring" <pri@xxxxxx>
Date: Fri, 5 Feb 1999 00:58:00 +0100
Importance: Normal
In-Reply-To: <Pine.GSO.3.96.981220220103.11730A-100000@gutenberg>
Sender: owner-framers@xxxxxxxxx
We made a few experiments the other day re. Acrobat Reader's search plugin and the case of search index file names on Solaris 2.5. The PDF documents and the search index catalog was created on Windows NT. The PDF documents have names such as '09006200.pdf'. All search index directories and files have uc names. The PDF documents was transferred to a ISO 9660 format cd-rom, which allows only uc A-Z, 0-9 and _ in 8.3 format in file names. That is, you cannot distinguish between uc and lc in file names on an ISO 9660 format cd-rom; unix'es often presents the file names in lc, but should find and and open a file whether you specify the name in uc or lc. What I found out was this: I can open the PDF documents and the attached search index directly from the cd-rom, which is mounted as a 'hsfs' file system. I suppose 'hsfs' reads High Sierra, which is another (older) name for the ISO 9660 standard. All file and directory names appear in _lc_ when I do a 'ls', but if I do 'ls /pathtocdrom/INDEX', the search index directory INDEX is listed. That is, the hsfs file system is not case sensitive. If I copy the PDF documents and the search index directory to a local directory, all files and directories are copied with lc names. And now the file system _do_ distinguish between uc and lc in file names. If an application, the Acrobat Reader search plugin in this case, is looking for the file INDEX.PDX or the directory INDEX, it won't find it. And Acrobat Reader cannot attach the search index. Now, if I rename the search index PDX file and the search index directory to uc, Acrobat Reader can attach the search index. Only those two was renamed, all PDF document file names and all the direcories and files below the search index directory remained in lc. What gives? Kind regards Peter Ring > -----Original Message----- > From: tmi@telelogic.se [mailto:tmi@telelogic.se] > Sent: 20. december 1998 22:06 > To: Peter Ring > Cc: FrameMaker user forum; framers@omsys.com > Subject: RE: Search index not available > > > On Fri, 18 Dec 1998, Peter Ring wrote: > > > > -----Original Message----- > > > > > > Apparently, the UNIX version of Acrobat Reader/Exchange > > > cannot read a search index whose filenames consist of only > > > uppercase letters. Acrobat Catalog creates the filenames > > > in uppercase, so we had to rename all the search index files > > > using only lowercase letters, and now the search index works! > > > > This does quite not correspond to what I found by searching the pdf-l > > archives at http://www.pdfzone.com. But I must admit that I > don't know what > > to conclude. It's just that I have to deliver a cross-platform > cd-rom with > > search index real soon ... > > I agree that the advice/solutions are confusing, but in our > case it was filenames in UPPERCASE being the culprit. This > is also what is says in an Adobe support document: > > > Acrobat Search running in SunOS cannot connect to or use indexes that > > are uppercase. To make the index available, convert all filenames of > > the index and its associated directory to lowercase. > > See http://www.adobe.com/supportservice/custsupport/SOLUTIONS/5aee.htm > > How this will affect your CD production is another matter... ** To unsubscribe, send a message to majordomo@omsys.com ** ** with "unsubscribe framers" (no quotes) in the body. **