CVS-Plugin uses the wrong encoding when reading the Entries files in a dual-boot system (windows/lin [message #318608] |
Wed, 01 August 2007 12:29 |
Mads Stavang Messages: 19 Registered: July 2009 |
Junior Member |
|
|
Hello,
I'll better start with the background information, or else my problem
won't make much sense:
I am trying to create a development environment which I can use on my
dual-boot system. I started with a Workspace created in Windows. So far
I done the following:
1. Mount a NTFS Partition under Linux. Check. Mounted as UTF-8 with ntfs-3g
2. Share Eclipse-Project between Linux/Windows. Check.
3. Reading filenames with non-ascii characters both with Windows-Eclipse
and Linux with Eclipse. Check.
4. Reading files encoded in ISO-8859-1 both with Windows-Eclipse and
Linux-Eclipse. Check.
5. Synchronizing CVS repository under Linux. Failed.
The reason why 5 failed is that CVS is using UTF-8 to read the
Entries-files (in the CVS directory), but those files were created under
Windows with cp-1252. If I commited a file "testäöü.txt" in Windows-CVS,
it would gain an entry in "Entries", encoded in cp-1252. But, if I
synchronize the same project afterwards in Linux (no changes made), then
a merging-conflict is announced, because Linux-CVS cannot find the
"testäöü.txt" in "Entries", instead it finds "test?.txt" (where ? is
some garbage character)
Even though I have defined ISO-8859-1 as default encoding for the
Workspace, this only applies to the "normal" files, not the CVS files.
I have tried to set LC_ALL=en_US.ISO-8859-1, but then the filenames are
affected also (item 3 in the list above).
Any suggestion how I can solve this?
-Mads
|
|
|
Re: CVS-Plugin uses the wrong encoding when reading the Entries files in a dual-boot system (windows [message #318624 is a reply to message #318608] |
Wed, 01 August 2007 13:37 |
Michael Valenta Messages: 560 Registered: July 2009 |
Senior Member |
|
|
Mads,
This sounds like a bug. Please log a bug report against Platform/CVS.
Michael
Mads wrote:
> Hello,
>
> I'll better start with the background information, or else my problem
> won't make much sense:
>
> I am trying to create a development environment which I can use on my
> dual-boot system. I started with a Workspace created in Windows. So far
> I done the following:
>
> 1. Mount a NTFS Partition under Linux. Check. Mounted as UTF-8 with ntfs-3g
> 2. Share Eclipse-Project between Linux/Windows. Check.
> 3. Reading filenames with non-ascii characters both with Windows-Eclipse
> and Linux with Eclipse. Check.
> 4. Reading files encoded in ISO-8859-1 both with Windows-Eclipse and
> Linux-Eclipse. Check.
> 5. Synchronizing CVS repository under Linux. Failed.
>
> The reason why 5 failed is that CVS is using UTF-8 to read the
> Entries-files (in the CVS directory), but those files were created under
> Windows with cp-1252. If I commited a file "testäöü.txt" in Windows-CVS,
> it would gain an entry in "Entries", encoded in cp-1252. But, if I
> synchronize the same project afterwards in Linux (no changes made), then
> a merging-conflict is announced, because Linux-CVS cannot find the
> "testäöü.txt" in "Entries", instead it finds "test?.txt" (where ? is
> some garbage character)
>
> Even though I have defined ISO-8859-1 as default encoding for the
> Workspace, this only applies to the "normal" files, not the CVS files.
>
> I have tried to set LC_ALL=en_US.ISO-8859-1, but then the filenames are
> affected also (item 3 in the list above).
>
> Any suggestion how I can solve this?
>
> -Mads
|
|
|
Re: CVS-Plugin uses the wrong encoding when reading the Entries files in a dual-boot system (windows [message #318664 is a reply to message #318624] |
Thu, 02 August 2007 08:45 |
Mads Stavang Messages: 19 Registered: July 2009 |
Junior Member |
|
|
:/
Bug reported:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=198621
Mads
Michael Valenta wrote:
> Mads,
>
> This sounds like a bug. Please log a bug report against Platform/CVS.
>
> Michael
>
> Mads wrote:
>> Hello,
>>
>> I'll better start with the background information, or else my problem
>> won't make much sense:
>>
>> I am trying to create a development environment which I can use on my
>> dual-boot system. I started with a Workspace created in Windows. So far
>> I done the following:
>>
>> 1. Mount a NTFS Partition under Linux. Check. Mounted as UTF-8 with
>> ntfs-3g
>> 2. Share Eclipse-Project between Linux/Windows. Check.
>> 3. Reading filenames with non-ascii characters both with Windows-Eclipse
>> and Linux with Eclipse. Check.
>> 4. Reading files encoded in ISO-8859-1 both with Windows-Eclipse and
>> Linux-Eclipse. Check.
>> 5. Synchronizing CVS repository under Linux. Failed.
>>
>> The reason why 5 failed is that CVS is using UTF-8 to read the
>> Entries-files (in the CVS directory), but those files were created under
>> Windows with cp-1252. If I commited a file "testäöü.txt" in Windows-CVS,
>> it would gain an entry in "Entries", encoded in cp-1252. But, if I
>> synchronize the same project afterwards in Linux (no changes made), then
>> a merging-conflict is announced, because Linux-CVS cannot find the
>> "testäöü.txt" in "Entries", instead it finds "test?.txt" (where ? is
>> some garbage character)
>>
>> Even though I have defined ISO-8859-1 as default encoding for the
>> Workspace, this only applies to the "normal" files, not the CVS files.
>>
>> I have tried to set LC_ALL=en_US.ISO-8859-1, but then the filenames are
>> affected also (item 3 in the list above).
>>
>> Any suggestion how I can solve this?
>>
>> -Mads
|
|
|
Powered by
FUDForum. Page generated in 0.02944 seconds