bug-lilypond
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: UNC file names


From: James Lowe
Subject: Re: UNC file names
Date: Fri, 10 Aug 2018 07:39:06 +0100
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.9.1

Hello,


On 09/08/18 19:02, Aaron Hill wrote:
On 2018-08-09 08:41, James Lowe wrote:
Sorry this took so long for me to get back to you.

More research tells me that it is not lilypond that is at fault here
but Windows.

Windows cmd does not support UNC paths.

That should not be relevant, though.  That at most limits the ability for the current working directory to be a UNC path, without first mapping it to a drive letter.  But it really should not affect the ability to invoke LilyPond and pass in a UNC path for the input file.

Possibly but I would guess that would only matter if you were running the command in cmd rather than a ps environment.

As an aside, PowerShell does not have the same working directory limitation, and you can `cd` to a UNC path as you wish.

Yes this is true and also of note I could not get LP to use UNC paths even when run from Powershell, so perhaps your previous paragraph is making this point?


But back to the issue, if LilyPond is ultimately calling CreateFile passing in the file path as specified in the command-line arguments, it should be able to open a UNC-based path providing there are no permissions issues.  What I would suspect is some quirkiness with MinGW/MSYS and Posix paths such that LilyPond is not generating the correct API call.

Yes, although that is well beyond my ken.


As such, what would be interesting is to get a Process Monitor capture of the failing case.  That way, we can see which specific file I/O calls failed and with which errors.  Unfortunately, I no longer use the Windows version of LilyPond, so I cannot immediately test this on my setup without having to set up a VM first.  If it is possible to run LilyPond in a portable mode without installation, then that would save significant time getting a test environment.

I can do that perhaps - although I haven't used proc mon for a while.

James



reply via email to

[Prev in Thread] Current Thread [Next in Thread]