[KDE/Mac] Review Request 127822: address the potential name/case clash of the Attica/Attica and Attica/attica header dirs

[Shaheed Haque]

AFAIK, this affects NTFS too. I believe the semantics are "store as
case-sensitive, retrieve as case insensitive".

And, FWIW, this was right PITA for the Python binding effort; in the
end, in my "unofficial" version I went with only shipping the new
CamelCase files where possible. I'm not exactly sure what solution
Stephen went with for the "official" CMake-driven version.


On 5 May 2016 at 08:48, René J.V. <rjvbertin at gmail.com> wrote:
> On Wednesday May 04 2016 22:51:04 Marko Käning wrote:
>
> Hi Marko,
>
>>a directory KF5/Foo/foo and the corresponding "CamelCase" headers in KF5/Foo/Foo .
>>
>>oh, no! Not again.
>
> Oh yes, "all of this has happened before and all of this will happen again" ...
>
>>I am case-insensitive here as well. This should not happen on such file systems!!!
>
> Well, it *will* happen on such filesystems. The only alternative is that the filesystem would raise an error when trying to create a new dir entry with a supposedly different name that maps to the name of an already existent entry. I think that's what higher-level applications like the Finder or the MSWin Explorer do, but it would probably break too many expectations for lower-level, Unix/Posix calls.
> The only solution is to avoid naming schemes where this kind of clash can occur. That is, to follow guidelines Michael Pyne posted the other day:
>
> 1. Do not accidentally lose information about differences in case (e.g. if
> user does "Save As" as "xSa" then the resulting file name sent to FS should be
> "xSa")
>
> 2. Do not rely on file names that differ only in case (e.g. don't install a
> fileA.foo and FileA.foo and expect to be able to open a precise one of those
> two later)
>
>>> The sheer amount of KF5 frameworks concerned means (IMVHO) that we should really strive for an upstream solution rather than having to maintain our own set of patches …
>>
>>I haven’t tested all this lately, but I remember that I ran into trouble with this because of phonon in Januar 15.
>
> Ah, that's not a KF5 framework, but I'll have another look. First impression though is that no clashes can occur, at least in its current MacPorts installation layout.
>
> R.
> _______________________________________________
> Kde-buildsystem mailing list
> Kde-buildsystem at kde.org
> https://mail.kde.org/mailman/listinfo/kde-buildsystem