Menu

problem with getting gedcom import to work

Help
Anonymous
2003-12-21
2003-12-25
  • Anonymous

    Anonymous - 2003-12-21

    siteref is http://teshb.darkscape.net/phpgedview/
         I loaded phpgedview on my home Wintel PC - works great (it may work from outside - maybe not try http://207.172.84.21/phpgedview ). Now I'm trying to get it to work on a ?linux box at a hosting site.
         The import GEDCOM never gives me any status of the import (as it loops through the dataset) on the linux host, but works like a champ on the wintel version.

    Any suggestions?

     
    • Arne Eckmann

      Arne Eckmann - 2003-12-21

      Hi Bradley

      Apart from the fact that it looks like your gedcom are not imported http://teshb.darkscape.net/phpgedview/ works.

      http://207.172.84.21/phpgedview cannot be accessed from her where I sit i Denmark.

      You have two gedcoms, one named tesh.ged and one named TESH.ged - on windows systems these two filenames are one and the same, and I'm not sure if it is a good idea to have two gedcoms with virtually the same name.

      Your gedcoms charset are "1 CHAR IBM WINDOWS", and as long as you only has US-ASCII inside the gedcom it probably doesn't matter much, but when you have some foreign characters in the gedcom the preferred charset is UTF-8.

      Have you tried to import the gedcoms into the database?

      Select the admin menu > Select administration of Gedcoms > in the table use import links to import the gedcoms to your database.

      best regard

      Arne

       
      • Anonymous

        Anonymous - 2003-12-21

        Arne,
             Thanks. The problem is exactly importing the gedcom - it does not seem to work. BTW, I've set up an id/pw as demo/demo.
             On the Windows version, when I import the GEDCOM, I get a bunch of status messages - a couple of screens worth - saying that individuals have been imported and then the all-important click here to continue. I don't get any of that on the Linux version: my little Mozilla lizard just spins for about 10-15 seconds, then puts me back to where I was.  And, yes, I know that TESH and tesh are the same in Windows, but the GEDCOM worked there first time, so I never needed to create any GEDCOMs beyond tesh.ged.
             Here's where I think the problem is: on the Windows version, tesh.ged_index.php is a 1.1k file. On the Linux version, tesh.ged_index.php is a 0 byte file. The Linux file props are set to owner/group 99, which I think is httpd or php. Rights are set to rwxr--r--. So, the index file is getting created, but not populated.

        Here's the listing from the rest of the ./index directory:

        drwsrwxrwx    3 teshb    users        4096 Dec 21 11:42 .
        drwxrwxrwx    8 teshb    users        4096 Dec 21 06:30 ..
        -rw-r--r--    1 99       99         208928 Dec 21 11:42 GENDEX_TESH.ged.txt
        -rw-r--r--    1 99       99         208928 Dec 21 11:41 GENDEX_tesh.ged.txt
        -rw-r--r--    1 teshb    users      555084 Dec 21 06:28 TESH.ged
        -rw-r--r--    1 99       99           7845 Dec 21 06:29 TESH.ged_conf.php
        -rw-r--r--    1 99       99              0 Dec 21 06:29 TESH.ged_index.php
        drwxr-xr-x    2 teshb    users        4096 Dec 21 11:39 old
        -rw-r--r--    1 99       99            281 Dec 21 11:33 pgv-20031221.log
        -rw-------    1 99       99           8011 Dec 21 11:42 sess_29be007d4dc35a7832c
        c5bf6a6123e6b
        -rw-------    1 99       99             56 Dec 21 10:37 sess_6b93653e49e6c40d574
        fae847d4489ef
        -rw-------    1 99       99             57 Dec 21 10:59 sess_8568ea2e9933398cd27
        1a9c130243cde
        -rw-------    1 99       99             56 Dec 21 08:54 sess_c72bf78c212386fc4cc
        0167ca9cf77c9
        -rw-------    1 99       99             56 Dec 21 08:18 sess_c7a8c9e6a24382df3ae
        07db38ddc37fd
        -rw-------    1 99       99             57 Dec 21 11:38 sess_e2d43cfec5afe195d3e
        ba8b95724c63c
        -rw-------    1 99       99             56 Dec 21 08:06 sess_efbe8ab816a59878e42
        c2e3060d16ea0
        -rw-rw-rw-    1 teshb    users      589290 Dec 20 14:43 tesh.ged
             After import, I get the usual error message mentioned in a lot of posts here (yup, read through most of them):
        Warning: Invalid argument supplied for foreach() in /home/teshb/public_html/phpGedView/functions_index.php on line 146

        Warning: Cannot modify header information - headers already sent by (output started at /home/teshb/public_html/phpGedView/functions_index.php:146) in /home/teshb/public_html/phpGedView/functions_print.php on line 492

        Do you want to take a stab at running the import? The GED is in http://teshb.darkscape.net/phpgedview/index/tesh.ged or TESH.ged. Feel free to delete whichever one you cannot re-import. As noted, the lowercase one is the downloaded one from the windows install - no foreign characters in there, I just accepted the default.

             In the editconfig_gedcom.php screen, I've goofed with increasing the timeout, but while it makes the import process run only slightly longer, it still doesn't make the ged_index file any bigger than 0.

         
    • Arne Eckmann

      Arne Eckmann - 2003-12-21

      Hi Bradley

      I tried to login and tried to upload the presidents.ged, but it stopped during the upload process: I tried it several times, but all in vain ;-(

      Could you please try to chmod the index folder with 777 - and the authenticate.php file with 777 and the config file with 777. In case your FTP don't understand 777 it is the most liberal setting you can have, and we can always lower it afterwards.

      Are the php files straight from the package, or have you by changed opened some of them in a editor on the windows platform ?

      The headers already sent message could eventually be caused by some unintentional whitespaces and unfortunately some editors on the Windows platform (including Notepad) has the nasty habit of adding (invisible) bytes at the start of files where it finds UTF-8 specific characters ?

      But first of all: can you try to set the most liberal rights (chmod) at the index folder, the authenticate.php and config.php.

      best regard
      Arne
       

       
    • Arne Eckmann

      Arne Eckmann - 2003-12-21

      Hi again Bradley

      I also tried to temporary placing your tesh.ged file at a similar Apache server, where it uploaded, imported and runs okay:

      http://phpgedview.1go.dk/index.php?GEDCOM=tesh.ged

      There are no problems with your gedcom file, and I can't see if there is a problem with your server, but I guess it's okay to - this leaves us with two possibilities which must be investigated:

      1. if the rights at the index folder, the authenticate.php file and the config.php file are sufficient for the application.

      2. if there somehow slipped some unintentional whitespaces into some of the php files.

      best regard

      Arne

       
    • tales

      tales - 2003-12-22

      Could this be the same problem I was having with my server? Perhaps it was timing out the program when you tried to import the gedcom? See my post on memory problems. Just a thought.

       
    • Mark Hattam

      Mark Hattam - 2003-12-22

      The tesh.ged has a few little errors (or non-standard GEDCOM) constructions

      some edited down parts of the records are ...

      Burial date is unrecognisable
      0 @I151@ INDI
      1 NAME Clyde Edward /KEELING/
      1 BURI
      2 DATE SeeNotes
      2 PLAC Hidden Valley Country Club

      Birth date is unrecognisable
      0 @I191@ INDI
      1 NAME Pearl /FLINCHUM/
      1 BIRT
      2 DATE ? 16 DEC 1910

      Death date is unrecognisable
      0 @I296@ INDI
      1 NAME Jane Elizabeth /RICHEY/
      1 SEX F
      1 DEAT
      2 DATE aged 74 years
      2 PLAC Abbeville Co., SC

      This one is not connected to any other record
      0 @I344@ INDI
      1 NAME Samuel /RITCHIE/ Jr.

      Death date is unrecognisable
      0 @I410@ INDI
      1 NAME Ellis C. /ALLEN/
      1 DEAT
      2 DATE ? JAN 1970
      2 PLAC Pahokee, FL
      2 NOTE maybe

      This one is not connected to any other record
      0 @I493@ INDI
      1 NAME Sheila Gay /FARMER/
      1 SEX M
      1 BIRT
      2 DATE 6 JAN 1952

      Death date is unrecognisable
      0 @I599@ INDI
      1 NAME Susan /EASTER/
      1 SEX F
      1 DEAT
      2 DATE bet 1860/1870

      Birth date is unrecognisable
      0 @I600@ INDI
      1 NAME Michael /EASTER/ II
      1 BIRT
      2 DATE bet 1760/1762

      Death date is unrecognisable
      0 @I604@ INDI
      1 NAME Michael III /EASTER/
      1 DEAT
      2 DATE bet 1860/1870

      This record is not connected to any others
      0 @I648@ INDI
      1 NAME Margaret Ann /WILSON/
      1 SEX F

      Birth date is unrecognisable
      0 @I813@ INDI
      1 NAME Isabel /FAIN/
      1 BIRT
      2 DATE ? 1 FEB 1914
      2 PLAC Dandridge, TN

      Death date is unrecognisable
      0 @I908@ INDI
      1 NAME Leland Telford /OWEN/
      1 DEAT
      2 DATE ? APR 1965
      2 NOTE maybe

      Death date is unrecognisable
      0 @I930@ INDI
      1 NAME John Milton /BURNS/
      1 DEAT
      2 DATE ? SEP 1975
      2 PLAC Dwight, IL

      Birth date is unrecognisable
      0 @I961@ INDI
      1 NAME Baylor Banks /HICKMAN/
      1 BIRT
      2 DATE ? 23 MAY 1898
      2 PLAC KY

      This record is not connected to any other
      0 @I962@ INDI
      1 NAME Laura Jane Thomas /TELFORD/
      1 SEX F
      1 BIRT
      2 DATE 18 DEC 1853
      1 DEAT
      2 DATE 12 JUN 1891
      2 PLAC Gainesville, GA

      This record is not connected to any other
      0 @I1050@ INDI
      1 NAME Sidney /TELFORD/
      1 SEX M

      Death date unrecognisable
      0 @I1358@ INDI
      1 NAME Sandra Jean /GRAY/
      1 DEAT
      2 DATE ? 2 MAR 1991
      2 PLAC Trussville, AL

      Birth date unrecognisable
      0 @I1704@ INDI
      1 NAME Joy /SIMPSON/
      1 BIRT
      2 DATE 26 may 19??

      Birth date unrecognisable
      0 @I1711@ INDI
      1 NAME Julie /WRIGHT/
      1 SEX F
      1 BIRT
      2 DATE 17 Oct 19??

      Birth date unrecognisable
      0 @I1789@ INDI
      1 NAME Alice Faye /ZWACK/
      1 SEX F
      1 BIRT
      2 DATE 19 Dec 19??

      This record not connected to any other
      0 @I1849@ INDI
      1 NAME Clyde Milton /TESH/
      2 SOUR @S2@
      1 SEX M
      1 BIRT
      2 DATE 3 OCT 1905

      This child was born 14 years before his father was born
      0 @I691@ INDI
      1 NAME John Caldwell /CALHOUN/
      1 SEX M
      1 BIRT
      2 DATE 1782
      1 DEAT
      2 DATE 1850

      and a lot more like these unrecognisable dates.

      Also my GEDCOM checker throws up errors about NOTE.NOTE in just about every note record, for instance

      0 @N5@ NOTE
      1 NOTE 9 children

      and some NOTES like this one aren't connected to any other record
      0 @N2@ NOTE
      1 NOTE served in CSA for three years 29th VA Infantry
      2 CONT
      2 SOUR @S10@

       
    • Anonymous

      Anonymous - 2003-12-22

      Thanks to all - the problem remains unresolved, but I think we're getting closer. Next stop is my webhost.

      Here's what I've done:
      --dl'd the latest PAF, imported my existing GEDCOM, which stripped off those weird things the Generations put into my GEDCOM export many moons ago. Re-exported it to GEDCOM (as tesh.paf5-export.ged), uploaded that file & tried to import to phpgedview - no go;
      --on the theory that a second filtration of my GEDCOM would result in an even cleaner file, I re-imported tesh.paf5-export.ged (no errors, but why not?), re-exported that as tesh-test2-pafexport.ged, uploaded that to my webhost & tried to import again - again no-go;

           I'm kind of stuck on the fact that files I ftp upload to the webhost is owner/group teshb/users, while files that are created by phpgedview is owner/group 99/99, over which I no longer have control, since I'm neither that user or in that group. I was going to ask my webhost to chown those files to 99/users and chmod them to 666 or 777. This still smells/acts like a rights issue.

      I'll let you all know if that resolves the problem.
      Thanks for your attention to this.

      P.S. - I originally started with this webhoster because he had offered free space to some emerging sf.net project - I wanted to contract with someone that gives back. If any of you have good recommendations of webhosts that are generous with sf.net - truly an invaluable resource on the web - let me know.

      Also, I left the FW port off of my earlier link to the WIndows version. Should've been http://207.172.84.21:8000/phpgedview - I haven't been outside my own firewall since I set this hole in the FW, so it still might not work.

       
    • John Finlay

      John Finlay - 2003-12-23

      How big is your gedcom file?
      What verison of PHP and Apache are you running?

      --John

       
    • Anonymous

      Anonymous - 2003-12-24

      <b>Problem solved.</b>

           My ISP patiently explained to me the implications of using the flatfile indexing and suggested I try using the MySql indexing - once I did that, it worked just fine.

           Also believing that my PAF-cleaned GEDCOMs were not completely clean, I created some subsets of those files (Mom's family, Dad's family, Dad's Mother's family) thinking that maybe, just maybe, the resulting clean combined GEDCOM was too much for this to handle.
           The resolution symptoms I had were thus:
      --importing presidential into MySql worked good/fast;
      --importing huge maxi-combo GEDCOM in MySql did not work;
      --importing smaller chunks of GEDCOMs into MySql worked;

           The smaller chunks of family GEDCOMs are only slightly bigger files than the combined GEDCOM file - combined was 891k, one of the smaller chunks is 625k. None of these GEDCOMs (including the presidential one at 422k) worked under the flat file indexing, so there may be some kind of bug that needs to be addressed.
           However, under MySql, everything hums like it should.

           <b>The result is that it all works now.</b>

           Thanks to all for helping me figure this out - this server stuff is always bewildering to me (I do wires & routers in real life, not server stuff).

       
    • John Finlay

      John Finlay - 2003-12-24

      How well the import works depends on the server setup and the PHP configuration along with the server load.

      Trying to import a large Gedcom onto a slow server with high load generally will time out or run out of memory before it completes.

      --John

       
    • Arne Eckmann

      Arne Eckmann - 2003-12-25

      Back from Christmas at Langeland I can say again, that I successfully imported the gedcom to a similar Apache Server at:

      http://phpgedview.1go.dk/index.php?GEDCOM=tesh.ged

      But, since I have no right to display that gedcom, I had to remove it again before I left for Christmast at Langeland.

      When I imported the gedcom at the 1go.dk Apache Server, I had it running as a flatfile index-mode with no problems.

      The size of the gedcom are about the same size as the presidents.ged - actually: tesh.ged = 575 KB (589.290 byte) on the Windows platform.

      best regard

      Arne

       

Log in to post a comment.