marc export: page source code exported

[[e-mail]]

eg 3.3.4

On Vandelay MARC export, we noticed some very weird behaviour.

When exporting MARC records via a CSV file, for the very first time with MARC8 encoding, you got the webpage source code in the exported file. The subsequent exporting via CSV files will always get the source code, regardless MARC8 or UTF8 (used to work before trying MARC8).

If you export one record via bib id using either MARC8 or UTF8, you always get the correct record.

After you tried MARC8 from CSV and got the source code, then tried to export one record via bib id, you will always get the correct MARC records exported from CSV with either MARC8 or UTF8.

The problem seems with the very first exporting from CSV with MARC8.

[Rogan Hamby (rogan-hamby)]

I can confirm this and see it a 3.3 system.

[Rosie Le Faive (rlefaive)]

I can confirm this. EG 3.4.2

It also happens to me when i load a file with extension .txt instead of .csv.
Or if I fail to put something in "Starts at 0..."

I have also experienced this when asking for MARCXML from the start. Under the same conditions, (.csv file, UTF-8) I can sometimes get MARC21.

It does appear that exporting one record by bibid, then trying MARCXML again with a CSV, works.

[Bill Erickson (berick)]

Here's a fix:

https://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/berick/lp1850825-vand-export-csv-fix

The code failed to initialize the source of the record IDs to CSV on page load, so some critical values were not sent to the server during the export.

[Rogan Hamby (rogan-hamby)]

Pre-patch the export worked on Firefox but not Chrome for me. Post patch, worked on both so looks good to me. Sign off at user/rogan/lp1850825_export_bibs_not_html

[Galen Charlton (gmc)]

Pushed to master, rel_3_5, and rel_3_4. Thanks, Bill and Rogan!