Text: fix DistanceFieldGlyphCacheGLTest failing on recent Mesa.

Neither a driver bug nor something wrong with the code. It's just
that with a tight flush rectangle the target texture may have some random
garbage left around the edges, causing the comparison to fail, so it's
now explicitly cleared upfront.

Doesn't happen when running the offending test alone (because I suppose
the memory is coming fresh from the driver, being zeroed out for security
purposes), only when running after the others (where I suspect it's now
reusing previous partially-filled memory which it didn't before). Doesn't
happen on NVidia either.

src/Magnum/Text/Test/DistanceFieldGlyphCacheGLTest.cpp

@@ -230,6 +230,20 @@ void DistanceFieldGlyphCacheGLTest::setImage() {
     CORRADE_COMPARE(inputImage->size(), (Vector2i{256, 256}));
 
     DistanceFieldGlyphCacheGL cache{data.sourceSize, data.size, 32};
+
+    /* Clear the target texture to avoid random garbage getting in when the
+       data.flushRange isn't covering the whole output */
+    Containers::Array<char> zeros{ValueInit, data.size.product()*pixelFormatSize(cache.processedFormat())};
+    cache.texture().setSubImage(0, {},
+        /* On ES2, R8Unorm maps to Luminance, but here it's actually Red if
+           EXT_texture_rg is supported */
+        #if defined(MAGNUM_TARGET_GLES2) && !defined(MAGNUM_TARGET_WEBGL)
+        cache.processedFormat() == PixelFormat::R8Unorm ?
+            ImageView2D{GL::PixelFormat::Red, GL::PixelType::UnsignedByte, data.size, zeros} :
+        #endif
+            ImageView2D{cache.processedFormat(), data.size, zeros}
+    );
+
     Containers::StridedArrayView3D<const char> src = inputImage->pixels();
     /* Test also uploading under an offset. The cache might be three-component
        in some cases, slice the destination view to just the first component */