test binaries for Beta 5

[Lubomir I. Ivanov]

On 5 August 2014 16:55, Dirk Hohndel <dirk at hohndel.org> wrote:
> On Tue, Aug 05, 2014 at 12:18:20PM +0300, Lubomir I. Ivanov wrote:
>> On 5 August 2014 09:09, Dirk Hohndel <dirk at hohndel.org> wrote:
>> >
>> > We know that the 32bit/Qt4 Windows installer has printing issues - so
>> > that's not the main goal of testing :-)
>> >
>>
>> actually it should all work the same way as the Qt5 build, except the
>> table print is in raster.
>>
>> re: tmpfile()
>> i remember Miika wrote the initial prepare_dives_for_divelogs()
>> variant for the GTK version, but it used tmpnam() if i recall
>> correctly. then i changed it to use tmpfile()...something in these
>> lines.
>>
>> TBH, i didn't notice that the file goes in C:\ on Win7, simply because
>> i use an administrator/root account.
>> testing on WinXP SP2, it goes in .\ which is no better.
>
> I simply ripped all this out and reimplemented it to go to
> C:\Users\<yourname>\AppData\Local\Temp or something.
>
> But of course now that I say this I don't know if I would have needed to
> sing some magic incantations and burn the excrements of some MSFT
> employees in order for this to work with non-ASCII (or non-Latin, or
> non-something) characters in the localized path name components...
>
> Could you take a look at commit 3fd8e50044c7 ?
>

yes, i saw it.
in general, subsurface_fopen() should be used in case the temp path
has special chars. on Win32.
(patch attached)

lubomir
--
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Web-use-subsurface_fopen-for-non-ASCII-paths-on-Win3.patch
Type: application/octet-stream
Size: 1290 bytes
Desc: not available
URL: <http://lists.hohndel.org/pipermail/subsurface/attachments/20140805/7c3ab9a9/attachment.obj>