UnCrash v0.5 BETA

[Carlo Hogeveen]

 

UnCrash has reached its first beta, meaning it now supports the basic functionality to restore a TSE session after a Windows crash.

A restored TSE session contains the (possibly unsaved) versions of files as close as possible to as they were at the moment of the crash.

The end result is that we are ASAP up and running again.

Again, it is the first beta, so bugs are likelier.

 

https://ecarlo.nl/tse/index.html#UnCrash

 

 

 

[C.H. Fred]

Hi.  I haven't been on a machine to look at your code yet.  Am curious how you implemented it?  Big picture summary.

-- 

Rick C. Hodgin

--

---
You received this message because you are subscribed to the Google Groups "SemWare TSE Pro text editor" group.
To unsubscribe from this group and stop receiving emails from it, send an email to semware+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/semware/00be01d778c7%2476aa8270%2463ff8750%24%40ecarlo.nl.

[Carlo Hogeveen]

 

Sure,

 

Algorithm 1 (session) summary:

At the start of a TSE session add the session’s meta information to a day journal.

Check the day journals for TSE sessions that were started and not ended

and are not a running TSE session, and if so offer to ignore or restore them.

If there are no crashed TSE sessions then delete logging and backup files

that are older than yesterday.

At the end of a TSE session add the session’s meta information to a day journal.

 

Algorithm 2 (backup) summary:

Every N (configurable) minutes check TSE’s open files.

If a file has a new FileChanged() value since the last check,

add its changed meta information to a session journal

and create a new date-time-named backup of the file to a backup folder.

 

Algorithm 3 (restore) summary:

The user has selected a session to restore.

Process the appropriate session journal to determine which files were open

with what meta information, and restore them as such from the backup folder.

Meta information includes a file’s changed-state, binary mode and a checksum,

the latter so each backup file can be verified and skipped for an older version if corrupt,

which is appropriate in an environment where crashes are a risk.

To view this discussion on the web visit https://groups.google.com/d/msgid/semware/CAOtq5xr8LrYjRvbS2UY%2BXxZj6FLM3yXEyvu0%2BB4Vs84sCtaJqA%40mail.gmail.com.