Custom Query (420 matches)

Kathy discovered there was a need for the following (copied from her email):

"how we can go about generating image plus text for pdfs. currently you can have item file with images, or html text which is ugly. It would be good to have item file with images plus text. they you can view images but search on the extracted text. do we write our own pdf box program? or maybe icecite already does this??"

Below are all the commits, first to get the GLI applet to work again for GS3, and then to convert the GLI applet to a Java Web Start application instead.

Instructions on configuring and using it are in gli\README_WebGLI.txt, and will be added to the wiki.

http://trac.greenstone.org/changeset/32127

http://trac.greenstone.org/changeset/32135 - http://trac.greenstone.org/changeset/32142

http://trac.greenstone.org/changeset/32150 - http://trac.greenstone.org/changeset/32156 (bugfix of http://trac.greenstone.org/changeset/32152 not relevant)

GLI wasn't able to download from https URLs because there was no SSL support. Some other surrounding issues were identified and also needed fixing.

I didn't at first know there was no https support and that this was because wget hadn't been compiled with SSL. I only discovered this after the following two fixes, which allowed better debugging:

1. Adding shutdownhook for GS2 to ensure the GS2 server is stopped on Ctrl-C. Tested on Linux.

http://trac.greenstone.org/changeset/31813 - http://trac.greenstone.org/changeset/31815

2. http://trac.greenstone.org/changeset/31816

"Fix to getRedirectURL() being stuck forever if the proxy details are wrong. I still can't work out what my proxy settings should be to get downloading to work from a TSG-adminsitered machine, but at least now GLI doesn't hang requiring a ctrl-C to terminate it."

3. Compiling up wget on Unix with open SSL to support https URLs required upgrading our wget to a newer version (from 1.15 to 1.17.1) and also compiling up OpenSSL now, which we didn't do before. We also got a precompiled windows wget binary of 1.17.1, compiled with SSL support, upgrading from shipping a windows wget 1.11.x binary that didn't include SSL.

• GLI changes:

http://trac.greenstone.org/changeset/31820 - http://trac.greenstone.org/changeset/31823

• OpenSSL:

http://trac.greenstone.org/changeset/31824

http://trac.greenstone.org/changeset/31835 (for Mac)

• Wget upgrade and setting no-check-certificate using Wgetrc conf file, needed in newer wget versions compiled on Linux to be able to download from https urls that don't have valid site certificates:

http://trac.greenstone.org/changeset/31825 - http://trac.greenstone.org/changeset/31842 and http://trac.greenstone.org/changeset/31854

4. Fixes to get proxying to work and surrounding issues

Further modifications to GLI getRedirectURL()

• http://trac.greenstone.org/changeset/31843

New class URLConnectionManager to enable accessing web pages in "no-check-certificate" mode from Java code. Needed by getRedirectURL()

• getRedirectURL() phased out: http://trac.greenstone.org/changeset/31852

To get proxying to work on Windows and other issues:

http://trac.greenstone.org/changeset/31851

http://trac.greenstone.org/changeset/31853

http://trac.greenstone.org/changeset/31856, http://trac.greenstone.org/changeset/31857

5. no-check-certificate supported through GLI Preferences interface rather than wgetrc conf file:

http://trac.greenstone.org/changeset/31860

6. Changes to get helpful messages on connection failures to display.

Only for windows, proxy_on, proxy_port and proxy_host would be set (along with username and pwd), while on Linux none of these would be set as proxy settings were obtained by wget from env vars that were set up.

To get helpful messages to display, proxy_on, proxy_port and proxy_host needed to be setup regardless of the OS.

http://trac.greenstone.org/changeset/31861 - http://trac.greenstone.org/changeset/31865, http://trac.greenstone.org/changeset/31874 - http://trac.greenstone.org/changeset/31876

7. Shifting to Windows using proxy settings set up in the environment (as linux does) rather than passing proxy flags to wget.

http://trac.greenstone.org/changeset/31877, http://trac.greenstone.org/changeset/31878

8. Shifting from single to support for multiple proxy URLs: one each for HTTP, HTTPS and FTP.

http://trac.greenstone.org/changeset/31879 - http://trac.greenstone.org/changeset/31881

9. GLI informs user when existing proxy settings in environment (from before launching GLI) are overridden with custom proxy settings turned on and set up within GLI.

http://trac.greenstone.org/changeset/31882 - http://trac.greenstone.org/changeset/31884