If you switch on the flag saying "leave the temporary files" - eg while debugging, and you change the original html then the new html is not used because the system re-reads the temporary files used from last time.

Or - eg
I just ran a complete reimport after amending my site html and the first document is not refetched

/var/www/mysite/sites/default/files/imported/articles$ ls -ltr 6R46.html*
-rw-r--r-- 1 www-data www-data 10946 2010-10-01 18:53 6R46.html
-rw-r--r-- 1 www-data www-data 11101 2010-10-03 15:31 6R46.html.1-tidied.html
-rw-r--r-- 1 www-data www-data 10986 2010-10-03 15:31 6R46.html.raw_xhtml.html
-rw-r--r-- 1 www-data www-data 68885 2010-10-03 15:31 6R46.html.5-data_structure.html
-rw-r--r-- 1 www-data www-data 11026 2010-10-03 15:31 6R46.html.3-after_templating.html
-rw-r--r-- 1 www-data www-data 10860 2010-10-03 15:31 6R46.html.2-after_rewriting.html

Comments

alexmc’s picture

alexmc commented

Am using Ubuntu Linux 10.4, PHP 5.3, Apache 2.x

dman’s picture

dman commented
Category: bug » feature

Yeah, that's pretty much by design.
The 'leave files' option is for deep debugging. I guess you have to note to delete them yourself in that case. Current debugging is working towards remote network spidering, so keeping copies speeds things up a lot on subsequent runs. Only the first step - the original source - is cached. The others are remade.

All cases I've worked on had static sources, and my repeat attempts were for tuning the template and options. I've not had a case where the source (static HTML) changed during an import.
I guess a "purge cache files" button may make sense.

alexmc’s picture

alexmc commented

I think what would be important is to say that if you switch this on then the original html files are not re-read.

In my situation I control the original html so I am trying to produce it in the semantic form that import_html expects. So it makes sense for me to both want to see the intermediate files, and refetch the html.

dman’s picture

dman commented
Status: Active » Closed (works as designed)

Clearing the old 6.x issues from the issue queue for a cleanup.