logs archiveBotHelp.net / Freenode / #ada / 2015 / July / 14 / 1
sparre
You could probably implement it without too much trouble, but I'm not sure how useful it would be.
charlie5
nor i tbh
was just thinking it might be nice to be able to review prior/next version of subprogs at the click of a button
darkestkhan
useless for new projects, but for older ones could be useful - new projects tend to have too quick moving API and too big/many changes for such feature to be feasible
charlie5
probably ... just a passing thought
sparre
darkestkhan: I would actually expect the feature to be more expensive on old projects, as rolling back through revisions until one finds a change to the selected section might take much longer.
Lucretia
charlie5: and fu*king hard to implement :D
x**e
darkestkhan
sparre: expensive yes, but on new projects it may well be totally useless
Lucretia
expensive AND useless
sparre
I would just implement the naively by rolling back one revision at a time checking first for changes to the relevant file and secondly to the relevant section.
Who would be interested in doing some work on gprbuild?
wtp
At work we use Beyond Compare and it is the best comparison tool we have used so far. However, despite the nice features there are with helping to fine tune the comparison (e.g. manually adjust which lines compare on each side), one thing I often think would be nice is if
the comparison tool knew how to process the computer language and be able to show that one function is the same but was simply moved.
Currently, when you have a few that move, it skrews up the entire comparison result and you basically have to manually check every single line and possible adjust
Being able to understand the language would make comparison results better.
darkestkhan
sparre: I would be
although I can't go to Copenhagen due to budgetary limits
sparre
wtp: That would require some kind of indexing of subprograms and types where you could search based on semantic similarity.
darkestkhan: Knowing who is interested is always a useful starting point. It may be that Copenhagen isn't the optimal location, and that we can get some sponsorship to cover a part of the costs.
darkestkhan
sparre: certain concept I recently entertained in my head is: have diff utility aware of PARTS of Ada syntax/semantics (mostly package/subprogram/declare blocks)
other than that it could be regular diff
wtp
I wonder if it would make sense for a Ada aware front end group up lines of code based on the language syntax and then feed the parts to an external diff to do the bulk of it. The front end could handle understanding such things as renamed types. moving of subprograms, and let the external diff handle the finer details
hmm, imagine the front end can be used to allow for navigating code while in the diff (e.g. hmm, I want to check definition of a particular type)
anyhow, it's easy to come up with lots of crazy ideas...
Lucretia
Why can I have SDL.C_Pointers (which is private package) and I can "private with" that inside SDL.Video.Surfaces but not inside SDL.Video.Surfaces.Makers?
sparre
Sounds like a compiler bug to me.
(or a programmer not doing what he thinks he's doing ;-)
Lucretia
private with SDL.C_Pointers;
package SDL.Video.Surfaces.Makers is
function Make (S : in SDL.C_Pointers.Surface_Pointer) return Surface;
end SDL.Video.Surfaces.Makers;
sdl-video-surfaces-makers.ads:4:30: missing "with SDL.C_Pointers;"
but
private package SDL.C_Pointers is ...
and I tested it with
http://pastebin.com/RZZa1epY << and that compiles
it's the exact same structure I have in SDL
well, that version with A.Bar doesn't compile, but if you compile a.c.d.e
next »