Developer forums (C::B DEVELOPMENT STRICTLY!) > Plugins development

AutoVersioning Plugin

<< < (26/41) > >>

foldingBLACKlinesSUCK:
Referring to Plugin included in CodeBlocks 8.02 (official release NOT SVN)


--- Quote ---make.exe: Nothing to be done for `all'.
Process terminated with status 0 (0 minutes, 0 seconds)
0 errors, 0 warnings

--- End quote ---


1) Why does Build Count increment even if BUILD is NOT DONE as in above message? Could you please fix it? :(
    It makes sense that it increments if "Rebuild" is done / "Clean" + "Build" but not the above case.

2) Why does the Revision increment randomly. The use? ofcourse, we can choose not to use it. But still, isn't it redundant?

GREAT PLUGIN other than that  :P

JGM:

--- Quote from: foldingBLACKlinesSUCK on March 23, 2008, 03:38:37 am ---1) Why does Build Count increment even if BUILD is NOT DONE as in above message? Could you please fix it? :(
    It makes sense that it increments if "Rebuild" is done / "Clean" + "Build" but not the above case.

--- End quote ---

Well, build count is a value to keep track of every test we make, to know how many times we ran the program no matter what (good for our bosses, if any). For the exact times that you build the program you should use the BUILD variable not the BUILDS_COUNT


--- Quote from: foldingBLACKlinesSUCK on March 23, 2008, 03:38:37 am ---2) Why does the Revision increment randomly. The use? ofcourse, we can choose not to use it. But still, isn't it redundant?

--- End quote ---

That was a suggestion of Thomas if I recall correctly, and is similar in the way microsoft does it on their IDE's. You can set the incrementation value of random numbers to 1 and will eventually increment by one and not randomly (just set preferences on the scheme tab).


--- Quote from: thomas on June 28, 2007, 11:47:52 am ---However, in some circles, it is also understood as:

* major: must be a bigger number than the competitor's product
* minor: increments with every new feature
* release: increments by a random amount in regular intervals, even in absence of changes to code
--- End quote ---

There I found it, release is equivalent to revision on the autoversioning plugin

mariocup:
Hi JGM,

i found a strange bug in your plugin.

How to reproduce:
1. Save two projects in a workspace
2. Close the workspace
3. Open a workspace in Code::Blocks that contains 2 projects
4. Select Project->Autoversioning

Now Autoversioning tries to configure the second project although the first project is activated. The strange thing is that if you open these projects (not from a workspace) everything works well.

Bye,

Mario

JGM:
If I recall correctly that was a problem of codeblocks not notifying the activated project after a workspace is completely loaded. First it reads the workspace then activates the first project so autoversioning sets it as the activated project, then loads the seconds projects and activates it so autoversioning change to this project as the activated one. After all the projects are loaded codeblocks switch back to the first project as the activated one but the plugin doesn't get notified so here is why the plugin still manages the seconds project (last one) and not the activated one (first one).

Anyways I will make some deep test this night if I have the time, since now I'm self employed and I'm working really hard to get my company going on. Thanks for pointing the problem  :D

JGM:
Well I finally had the time to check the codeblocks sources, since I want to be able to create a patch ready to applied to svn. Well It has been a lot of time I doesn't saw c++ code, actually like 3 weeks, that's for me too many time :shock:

Well, mariocup  this is the event that was not getting notified the last time I worked with the plugin, with the situation that you described.


--- Code: ---void AutoVersioning::OnProjectActivated(CodeBlocksEvent& event)
{
if (IsAttached())
{
// switch to the settings of the now activated project [I assume it has already been loaded before this triggers]
m_Project = event.GetProject();
}
}

--- End code ---

I haven't found the time to make another deep test since only I had the time to check the new sources and write this message, but I think is a codeblocks problem, like the other similar problem that I found with the oncompilerstartedevent that was fixed on codeblocks some hundred of revisions ago :D But when I make the deep test I will report back so developers could fix it if possible.

PS: That nick "mariocup" reminds me of the nintendo 64 game mario kart :) oohh those days!!

Navigation

[0] Message Index

[#] Next page

[*] Previous page