Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    jl
    @jlillest
    I make chat good
    Andy Edwards
    @jedwards1211
    Man should I start using this instead of e-mail and sporadic GitHub comments for discussions?
    vpicaver
    @vpicaver
    Yea, sounds good
    Andy Edwards
    @jedwards1211
    so I was thinking about how to do data merging
    I think in general we can just look for shots with the same exact combo of from/to/dist/azm/inc
    if two trips have just one match like that, we assume they're the same
    otherwise we let the user decide
    and when we "merge" one trip into another, we can actually just replace it with the entire new trip data, so that deleted shots will stay deleted after the "merge"
    also for trips that are new, we can see if they're siblings of matched trips in the tree and if so, mark them to add to the matched trips' cave by default
    do you see any problems with that approach?
    vpicaver
    @vpicaver
    So have thought about this a lot in the pass
    I really want to have something where multiple people could work on the file and merge it together
    It might be interesting to have a mergetool if using version control.
    One idea I had was using uuid for for each object. This won't work very for merging from walls, compass, survex
    vpicaver
    @vpicaver
    We also need to figure out how handle merge conflicts
    vpicaver
    @vpicaver
    We might be able to use qt's meta system to help use do the iteration for merge comparison.
    For non qobjects we could use qgadget for simple types
    It may reduce code maintain, new type and data could be merged automatically
    vpicaver
    @vpicaver
    We may need to have an ignore property list for each type if we don't want to compare the properties
    If using the meta system
    The hard bit is how do we merge slightly modified lists
    I haven't figured that out yet
    Andy Edwards
    @jedwards1211
    Huh. Yeah, I can see how that would make sense for multiple people editing data in cavewhere directly
    one question about using uuids -- what happens if two people independently add the same shots?
    my focus was different, if people are using walls for instance it will probably remain their primary data source, and there just needs to be a smooth process to import updated walls data into cavewhere
    the thing that makes that tricky is people could change cave, trip, and station names
    or also station prefixes, which might change a whole slew of station names at once
    but for instance, if someone corrects a measurement or two in a trip -- most of the other shots remain the same, and if just one matches the station names and measurements of a shot already in Cavewhere, there's a pretty good chance it's the same trip, even if the trip was renamed
    Andy Edwards
    @jedwards1211
    optionally I could use a levenshtein distance tree to look for the most similar trip names in case typos are corrected, but for complete renames matching measurements is probably the best bet
    I haven't really analyzed it but I'm assuming shots with the same exact measurements are pretty rare...of course shots with the same from/to station will be even rarer except in the case of resurvey...
    but yeah what I'm describing wouldn't work well in the case of multiple changes by different people in different places...it would only work for updating cavewhere data
    I'm sure git works well for managing changes to survex, compass, and walls data files, if we eventually had metacave it might work well for that too
    Andy Edwards
    @jedwards1211
    provided we format the metacave JSON with plenty of line breaks
    vpicaver
    @vpicaver
    We could possible write our own get merge tool.
    It would be hard to detect the instance where people have enter duplicate data with uuid method.
    vpicaver
    @vpicaver
    With the uuid method you would have to worry about renaming issues. Although it would add 16 byte of data to each object
    You wouldn't have to worry about renaming issues
    I really want cave where to get converted to meta cave so it can easily be added to version control
    Julian Kirsch
    @ThisIsJulian
    Hey there,
    I've a question concerning the use of Qt Build Suite in your project.
    Do you integrate any code analysis system with Qbs?
    vpicaver
    @vpicaver
    We currently do not...
    What kind of code analysis system are you trying to integrate into qbs?
    Julian Kirsch
    @ThisIsJulian
    I want to integrate 'SourceMonitor' for Code metrics.
    But the choice of tool is optional. Few colleagues suggested that I should use one for my project :)
    So, I am basicly looking how other people solve this in their projects