regression testing: page image differences
Here are output differences between 20070112 and 20070123, mostly to test the regression tests. Here's how to decode it. The line WARN: tex/context/base/math-pln.tex/test-001-20070112-20070123.report means tex/context/base/math-pln.tex/test-001.tex was run through 20070112 and 20070123 to make pdf files, converted to bitmaps, and subtracted (page by page), and that there were some differences. The line [p.0001] 0.176462 dB diff [see png's in /tmp/tmp.mRMxS24124/] menas that p.1 was different in the two runs, but only slightly (about 0.2 dB). Here another example with a slightly larger dB difference: WARN: tex/context/base/core-mat.tex/test-001-20070112-20070123.report [p.0001] 5.00882 dB diff [see png's in /tmp/tmp.UllhY25247/] for which I've attached the two images and their difference. It's slight character position shifts, I think. WARN: tex/context/base/math-pln.tex/test-001-20070112-20070123.report [p.0001] 0.176462 dB diff [see png's in /tmp/tmp.mRMxS24124/] WARN: tex/context/base/unic-034.tex/test-001-20070112-20070123.report [p.0001] 1.68318 dB diff [see png's in /tmp/tmp.lkiMH24898/] WARN: tex/context/base/unic-039.tex/test-001-20070112-20070123.report [p.0001] 0.0678701 dB diff [see png's in /tmp/tmp.QRvMD25106/] WARN: tex/context/base/core-mat.tex/test-004-20070112-20070123.report [p.0001] 4.56087 dB diff [see png's in /tmp/tmp.jtiqZ25156/] [p.0002] 4.83235 dB diff [see png's in /tmp/tmp.jtiqZ25156/] WARN: tex/context/base/core-mat.tex/test-001-20070112-20070123.report [p.0001] 5.00882 dB diff [see png's in /tmp/tmp.UllhY25247/] WARN: tex/context/base/core-mat.tex/test-002-20070112-20070123.report [p.0001] 3.31206 dB diff [see png's in /tmp/tmp.TFjDk25293/] WARN: tex/context/base/core-mat.tex/test-003-20070112-20070123.report [p.0001] 0.475091 dB diff [see png's in /tmp/tmp.uaWCt25342/] WARN: tex/context/base/unic-000.tex/test-001-20070112-20070123.report [p.0001] 0.0814442 dB diff [see png's in /tmp/tmp.BDorO26317/] WARN: tex/context/base/unic-003.tex/test-001-20070112-20070123.report [p.0001] 0.0678701 dB diff [see png's in /tmp/tmp.lzEhB26435/] WARN: tex/context/base/core-itm.tex/test-001-20070112-20070123.report [p.0001] 0.352925 dB diff [see png's in /tmp/tmp.RDqzg26640/] [p.0002] 0.746572 dB diff [see png's in /tmp/tmp.RDqzg26640/] [p.0003] 1.85964 dB diff [see png's in /tmp/tmp.RDqzg26640/] [p.0004] 4.57445 dB diff [see png's in /tmp/tmp.RDqzg26640/] [p.0005] 1.03163 dB diff [see png's in /tmp/tmp.RDqzg26640/] -Sanjoy `A nation of slaves is always prepared to applaud the clemency of their master who, in the abuse of absolute power, does not proceed to the last extremes of injustice and oppression.' (Gibbon)
On Mon, 5 Feb 2007, Sanjoy Mahajan wrote:
Here are output differences between 20070112 and 20070123, mostly to test the regression tests. Here's how to decode it. The line
WARN: tex/context/base/math-pln.tex/test-001-20070112-20070123.report
means tex/context/base/math-pln.tex/test-001.tex was run through 20070112 and 20070123 to make pdf files, converted to bitmaps, and subtracted (page by page), and that there were some differences. The line
[p.0001] 0.176462 dB diff [see png's in /tmp/tmp.mRMxS24124/]
menas that p.1 was different in the two runs, but only slightly (about 0.2 dB). Here another example with a slightly larger dB difference:
Could it be because of the differences between lm-math and cm-math fonts? Aditya
menas that p.1 was different in the two runs, but only slightly (about 0.2 dB). Here another example with a slightly larger dB difference:
Could it be because of the differences between lm-math and cm-math fonts?
Good thought. I'm pretty sure you're right. Here's one of the 5dB differences. Each pdf file uses five fonts, and the two fonts that vary between the files are the math ones: $ pdffonts tex/context/base/core-mat.tex/test-001-20070112.pdf > 1 $ pdffonts tex/context/base/core-mat.tex/test-001-20070123.pdf > 2 $ diff -U0 1 2 --- 1 2007-02-05 21:29:55.000000000 -0500 +++ 2 2007-02-05 21:29:55.000000000 -0500 @@ -4 +4 @@ -XXFRMY+CMMI12 Type 1 yes yes yes 5 0 +URGTRZ+LMMathItalic12-Italic Type 1 yes yes yes 5 0 @@ -7 +7 @@ -FXXUVH+CMSY10 Type 1 yes yes yes 8 0 +DOUOBX+LMMathSymbols10-Italic Type 1 yes yes yes 8 0 -Sanjoy `Not all those who wander are lost.' (J.R.R. Tolkien)
Sanjoy Mahajan wrote:
menas that p.1 was different in the two runs, but only slightly (about 0.2 dB). Here another example with a slightly larger dB difference:
Could it be because of the differences between lm-math and cm-math fonts?
Good thought. I'm pretty sure you're right. Here's one of the 5dB differences. Each pdf file uses five fonts, and the two fonts that vary between the files are the math ones:
That is my guess as well (looks like a hinting difference, actually). Best, Taco
participants (3)
-
Aditya Mahajan -
Sanjoy Mahajan -
Taco Hoekwater