Home » Community » U++ community news and announcements » Use unique output directory per assembly (append assembly name to output director)
|
|
|
|
| Re: Use unique output directory per assembly (append assembly name to output director) [message #31493 is a reply to message #31490] |
Sun, 06 March 2011 22:08   |
|
|
Good idea Mirek!
I understand your motivation very well... I have some asemblies that differ just in one directory, e.g. one works with stable sources and the other with development version. Few times I got terribly stuck with weird bugs, which I later discovered to be caused by out-of-date .o files. If I understand correctly, this should solve this in future.
I know you considered using md5 or some other hash as the unique identifier, why did you decide to use name in the end? Actually I would prefer the unique part to be hash of the nests inside given assembly, even though the output directory names would be less readable. The reason why I would benefit from this is that I use two different ides ("theide", which is nightly build and "ide", which contains some personal tweaks), so that they would use the same output for assemblies that have same hashes, regardless of their name... On the other hand I understand that this is probably not very common setup and that I should probably use consistent naming (for my own good  ). So this is just a comment, without any real urge to change your decision 
Thank you for this feature,
Honza
|
|
|
|
|
|
|
|
| Re: Use unique output directory per assembly (append assembly name to output director) [message #31498 is a reply to message #31496] |
Sun, 06 March 2011 23:27 |
|
|
| mirek wrote on Sun, 06 March 2011 22:49 |
| dolik.rce wrote on Sun, 06 March 2011 16:08 |
I know you considered using md5 or some other hash as the unique identifier, why did you decide to use name in the end? Actually I would prefer the unique part to be hash of the nests inside given assembly, even though the output directory names would be less readable.
|
Well, mostly because assembly name is unique.
Sure, nests can change, but they can change "materially" too (like overwriting nest with other branch), so basically you still have to remember that after fiddling with it too much, you have to "bomb" it...
|
As I said, I have two theides, which have clashing assembly names. Right now I use separate output directories to solve that. Renaming the assemblies would be an option too, of course This is corner case, but the other advantages of nest based hashes (like detecting assembly renaming, same assemblies etc.) seemed quite compelling and useful to me.
Also overwriting the sources shouldn't be a problem. Unless you manage to keep older file timestamps
Honza
|
|
|
|
Goto Forum:
Current Time: Thu Apr 25 22:49:54 CEST 2024
Total time taken to generate the page: 2.94552 seconds
|
Members
Pages
Help
Register
Login
Home
