From a60a735bd015d045c4fdeceb8f26591918dd839d Mon Sep 17 00:00:00 2001 From: Ilya Kreymer Date: Tue, 17 Mar 2015 00:14:10 -0700 Subject: [PATCH] Update INSTALL.rst for 0.9.0 --- INSTALL.rst | 30 ++++++++++++++++++++++++++---- 1 file changed, 26 insertions(+), 4 deletions(-) diff --git a/INSTALL.rst b/INSTALL.rst index e5259b60..3e55236d 100644 --- a/INSTALL.rst +++ b/INSTALL.rst @@ -1,14 +1,16 @@ Installation ============ -This section covers more detailed installation info for pywb. +This section covers more detailed installation info for pywb. + +*These instructions apply to older versions of pywb +but will still work with pywb 0.9.0, although the directory based configuration system and ``wayback-manager`` utility +remove some of these steps.* Requirements ~~~~~~~~~~~~ -pywb has tested in python 2.6, 2.7 and pypy. - -It runs best in python 2.7 currently. +pywb has tested in python 2.6, 2.7. It runs best in python 2.7 currently. pywb tool suite provides several WSGI applications, which have been tested under *wsgiref* and *uWSGI*. @@ -157,3 +159,23 @@ For more advanced use, the pywb init path can be customized further: `wayback.py `_ and `pywb\_init.py `_ for examples of existing initialization paths. + + +A note on CDX index files +""""""""""""""""""""""""" + +The new ``wayback-manager`` tool will automatically generate CDX index files for all WARCs and ARCs, so +manual updating of CDX indexes is no longer required. + +The ``cdx-indexer`` also creates files in the `SURT `_ format by default. format + +However, if you need to use existing/legacy .cdx files, you may need to set a special config (for now). + +If you are using .cdx files where the key is *not* in SURT format (that is, the CDX line may start with ``example.com`` instaed of ``com,example)/``), +simply add the following to the main ``config.yaml`` +:: + + surt_ordered: false + +A SURT CDX key reverses the order of domain and subdomains and allows for improved searching. +Future versions of pywb may detect the format automatically.